Experiences of NHS healthcare services in England: July 2025

• The percentage of people who found it "difficult" to contact their GP practice decreased from 18.7% in Wave 1 (23 July to 15 August 2024) to 10.6% in Wave 12 (27 May to 18 June 2025).

• The percentage of people who had a "poor" overall experience of their GP practice decreased from 15% in Wave 1 (23 July to 15 August 2024) to 10.9% in Wave 12 (27 May to 18 June 2025).

• The estimates show that access to and satisfaction with GP practices varies by integrated care board (ICB).

• The extent to which the demographic and socio-economic characteristics of the ICB populations explain that this variation differs by ICB.

• Even after adjusting for demographic and socio-economic factors, variation remains between ICBs.

The Health Insight Survey (HIS) collects a wide range of data on individuals' self-reported experiences of NHS healthcare services in England. The survey is commissioned by NHS England and more information can be found on our About the Health Insight Survey (HIS) web page.

The Health Insight Survey team have published 12 waves of HIS data between September 2024 and July 2025, which is sufficient to undertake an analysis of trends. This bulletin looks at integrated care board (ICB) trends over time and explores whether differences in GP access and satisfaction between ICBs can be explained by underlying demographic and socio-economic differences in the population of each ICB.

This bulletin focusses on four survey questions, and data on the other questions are available in our accompanying datasets. Confidence intervals are provided for all estimates as an indication of the degree of uncertainty. These are wider for integrated care boards (ICBs) with a small sample size, therefore the estimate is less precise.

Figure 1: The percentage of adults finding it "difficult" to contact their GP practice has decreased over time

Estimated percentage of adults who found it "difficult" to contact their GP practice, by integrated care board, Wave 1 to Wave 12, England

Overall, the percentage who found it "difficult" or "very difficult" to contact their GP practice decreased from 18.7% in Wave 1 (23 July to 15 August 2024) to 10.6% in Wave 12 (27 May to 18 June 2025).

From Wave 1 to Wave 12, the percentage of people who found it "difficult" to contact their GP decreased in 41 ICBs.

Figure 2: The percentage of people waiting more than a day to know how their request will be dealt with has remained stable

Estimated percentages of people who waited more than a day to know how their request would be dealt with, or who never found out how their request would be dealt with, after contacting their GP practice, by ICB, Wave 1 to Wave 12, England

Overall, the percentage of people who waited more than a day to know how their request would be dealt with, or who never knew how it would be dealt with, has remained stable over all waves, changing from 29.6% in Wave 1 to 28.2% in Wave 12.

Figure 3: The percentage of people who are not able to see their preferred healthcare professional varies by ICB

Estimated percentage of people who were not able to see their preferred healthcare professional, by integrated care board, Wave 4 to Wave 12, England

Notes:

1. From Wave 4 (15 October to 16 November 2024), respondents with a preferred healthcare professional were asked if they were able to see them, data is not available before this.

Overall, of those who have a preferred healthcare professional, the percentage who were not able to see them decreased slightly from 39.0% in Wave 4 to 34.6% in Wave 12.

Figure 4: The percentage of people reporting a "poor" overall experience of their GP practice has decreased for most ICBs

Estimated percentages of people who reported an overall "poor" experience of their GP practice, by ICB, Wave 1 to Wave 12, England

Overall, the percentage of adults reporting a "poor" or "very poor" experience of their GP practice has decreased from 15.0% in Wave 1 to 10.9% in Wave 12.

The percentage of adults who reported a "poor" overall experience decreased in 38 ICBs from Wave 1 to Wave 12.

The published estimates show that access to GP practices, and satisfaction with GP practices, varies by integrated care board (ICB).

The data also show that access to GP practices, and satisfaction with GP practices, varies by core demographic and socio-economic (population) characteristics. For example, those in the younger age groups are more likely to report finding it "difficult" to contact their GP practice.

We undertook logistic regression modelling using combined data from Waves 7, 8, and 9 to give a larger sample size. The purpose of this analysis is not to rank ICBs but to explore whether some of the variation between ICBs can be explained by differences in the underlying population characteristics of each ICB.

We report the results for the question "How easy or difficult was it to make contact with your GP practice?". The model outputs for whether participants knew how soon their request would be dealt with, whether they were able to see their preferred healthcare professional, and whether they rated their overall experience as "good", or "poor", can be found in the data download for Figure 6. The findings were similar across all four questions.

Difficulty contacting a GP practice

Participants in NHS Herefordshire and Worcestershire were least likely to report it being "difficult" to contact their GP practice. We set this ICB as the baseline to explore the extent to which population characteristics explain the variation between ICBs.

The extent to which population characteristics explain the variation between NHS Herefordshire and Worcestershire and the other ICBs, in finding it "difficult" to contact a GP practice, differs by ICB.

Figure 5 shows the five ICBs where the greatest proportion of the variation to NHS Herefordshire and Worcestershire is explained by population characteristics. For two of these ICBs, NHS North West London and NHS Frimley, controlling for population characteristics means it is no longer significantly more "difficult" to contact a GP practice compared to NHS Herefordshire and Worcestershire. However, for three ICBs it is still significantly more "difficult" to contact a GP practice compared with NHS Herefordshire and Worcestershire after adjusting for population characteristics.  

Figure 5: The extent to which population characteristics explain the variation between ICBs differs by ICB

Odds ratios and confidence intervals for the baseline model and fully adjusted model for selected ICBs, Waves 7 to 9 combined, England

Notes:

1. The values in the dot chart are odds ratios; this is a comparison of the odds of finding it "difficult" to contact a GP practice between an individual integrated care board (ICB) and NHS Herefordshire and Worcestershire (the reference group).

2. This chart shows the odds ratios for the baseline model and fully adjusted model for the five ICBs where the greatest proportion of the variation to NHS Herefordshire and Worcestershire is explained by population characteristics.

After only adjusting for population characteristics that were available in the Health Insight Survey, statistically significant variation remains between ICBs (Figure 6). However, it is likely that unmeasured differences remain between the populations of each ICB, such as educational attainment, which could influence how people report GP access and satisfaction. Therefore, we cannot be certain that the remaining variation in ICB results is because of differences in operational practices, workforce, or other internal factors within the ICBs.

Figure 6: Variation remains between ICBs after adjusting for population characteristics

Odds ratios for integrated care boards (ICBs) for the fully adjusted model, Waves 7 to 9 combined, England

Integrated care board (ICB)

Integrated care boards (ICBs) are NHS organisations in England responsible for planning and managing local health services.

Logistic regression

Method of analysis used to predict the odds of a binary response variable using one or more explanatory variables.

Odds ratio

The ratio of the odds for a binary variable between two groups. For this analysis, it represents the odds of a respondent answering "yes" in one ICB compared with the odds of a respondent answering "yes" in another ICB. Odds ratios are useful for interpreting the associations between ICBs and responses; an odds ratio different from one suggests there is an association.

Proportion of variation

The proportion of variation between an ICB and the reference group that can be accounted for by demographic and socio-economic factors is calculated as the change in odds ratios between the baseline model and the fully adjusted model, relative to the baseline model odds ratio.

The time series figures show weighted estimates from Wave 1 to Wave 12 and are available in our accompanying datasets.

For the logistic regression modelling, we aimed to use recent data, and to also combine data across waves to give a larger sample size and reduce uncertainty in the estimates.

To achieve this, we grouped the published data in to sets of three waves; Waves 1, 2 and 3, Waves 4, 5 and 6, and Waves 7, 8 and 9. The unweighted data from the most recent complete set of three waves available at the time of analysis (Waves 7, 8, and 9) were combined.

We filtered the data to include only participants who had attempted to contact their GP practice in the past 28 days in one or more waves.

If a participant attempted to contact their GP practice in the past 28 days in more than one wave, we only kept data from the most recent attempt.

We ran three logistic regression models for each question:

• a baseline model including only integrated care board (ICB) as a predictor variable, to estimate the initial difference across ICBs before adjusting for other variables

• a demographic model including ICB and demographic variables (age, sex and ethnicity), to explore how much of the initial effect of ICB can be explained by demographic factors

• a fully adjusted model including ICB, demographic variables and long-term health condition, deprivation and urban or rural geography to explore how much of the initial effect of ICB can be explained

Predictor variables for each model

We only controlled for variables in the Health Insight Survey. Other variables, such as educational status or income, might explain the association between ICB and outcome.

We explored interactions between age and sex, and whether someone was contacting their GP practice for themselves or for someone else in their household. However, these predictors were excluded from the models, as they did not significantly affect the ICB odds ratios.

Baseline model predictor variables

ICB

We used the ICB with the lowest percentage of participants, selecting the outcome coded as "1" as the reference group for each model.

Mode of contact (question two only)

• telephone (reference group)

• in person

• online

• other

Demographic model predictor variables

The demographic model includes the predictor variables from the baseline model and three additional sets of predictor variables:

Age group:

• 16 to 24 years

• 25 to 34 years

• 35 to 44 years

• 45 to 54 years

• 55 to 64 years (reference group)

• 65 to 74 years

• 75 years and over

Sex:

• female (reference group)

• male

Ethnicity:

• A combined grouping of: "Asian or Asian British", "Black or African or Caribbean or Black British", "Mixed or Multiple Ethnic Groups" and "Other ethnic group"

• White (reference group)

Fully adjusted model predictor variables

The fully adjusted model includes the predictor variables from the baseline and demographic models and three additional sets of predictor variables:

Index of Multiple Deprivation (IMD):

• quintile 1 (most deprived)

• quintiles 2 to 5 (reference group)

Long-term health condition:

• "yes"

• "no" or "prefer not to say" (reference group)

Urban or rural geography:

• urban (reference group)

• rural

Bootstrapping

We used bootstrapping to determine confidence intervals for the differences in odds ratios between the baseline and fully adjusted models. We resampled the quarterly sample (Waves 7, 8, and 9) 10,000 times, fitted the baseline and fully adjusted models to each sample, and generated the odds ratios.

We constructed 95% confidence intervals for the difference in odds ratios between the two models for every ICB. There was a statistically significant difference in odds ratios between the baseline model and fully adjusted model for 31 out of the 41 ICBs (excluding the reference group).

More quality and methodology information

More quality and methodology information (QMI) on strengths, limitations, appropriate uses, and how the data were created is available in our Experiences of NHS healthcare services in England QMI.