Patient satisfaction with general practice in Scotland: secular trends and association with deprivation, rurality and practice size—a secondary analysis of national survey data

Abstract

Objectives

To explore the relative contributions of practice size, socioeconomic deprivation and rurality to patient experience in Scotland and to examine temporal trends in patient satisfaction.

Design

Secondary analysis of repeated cross-sectional national survey data.

Setting

Primary care in Scotland, using data from the Scottish Health and Care Experience (HACE) survey conducted biennially between 2009–2010 and 2021–2022, linked to publicly available national datasets.

Participants

Data from 632 general medical practices (63% of the original 1008 participant practices) were analysed. Practices were excluded for reasons including potential closures or mergers, changes in urban–rural classification or missing population size data. Response rates were 16%–38%, highest in the most remote and least deprived areas.

Primary and secondary outcome measures

Practice-level percentage positive responses to seven HACE survey items, relating to patient-centred care and continuity of care, were assessed based on questions where respondents were asked to indicate their level of agreement. The relative importance method and regression models were used to examine the contribution of eightfold urban-rural classification, area deprivation quintile and practice size quintile to models of patient satisfaction within the latest dataset.

Results

Satisfaction levels captured by most questions remained broadly constant until 2019-2020, but there was a marked decline in reported satisfaction in the 2021-2022 survey. Participants in the more rural areas and patients registered with smaller practices consistently reported higher satisfaction levels over time. Participants registered with practices in more affluent areas expressed moderately higher levels of satisfaction in most domains. The relative importance analysis emphasised the dominant role of socioeconomic status and practice size in influencing patient satisfaction across most survey items (area deprivation accounting for up to 71% and practice population for up to 86% of explained variance for individual items), while rurality made a smaller contribution (5%–18%). Findings were confirmed in sensitivity analyses, excluding large practice size changes and minor changes in survey wording.

Conclusions

Patient satisfaction underwent a marked decline after 2019–2020. Patients living in more rural and more affluent areas, and those registered with smaller practices, generally reported the highest levels of satisfaction. In the most recent survey, socioeconomic status and practice size were most strongly associated with patient satisfaction across most survey items. Further work is needed to clarify the mechanisms behind these findings.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• General practice in Scotland is publicly funded and free at the point of use to all citizens.

• Patients registered with all general practices in Scotland were sampled biennually.

• The methods employed in the national Health and Care Experience surveys have been largely consistent over time.

• Response rates to the surveys are relatively low and they vary significantly with demographic factors, with lower participation observed in urban and deprived areas.

• Variations in the survey questions over time necessitated the comparison of similar, rather than identical, questions prior to 2017–2018.

Introduction

Patient satisfaction with general practice is a useful indicator of service quality, reflecting the breadth of patient experience.¹ Overall satisfaction has been reported to be associated with personal continuity of care,^{2 3} patient socioeconomic status,¹ small practice size² and rurality,⁴ and there is evidence that continuity of care has declined in the UK over time.^{5 6} In this paper, we have conducted a secondary analysis of selected questions from national survey data to examine temporal trends in reported Scottish patient experience with general practice since 2009–2010 and to assess the relative importance of three independent variables, socioeconomic status, practice size and rurality—variables which are likely to be strongly correlated with each other. Given the evolving landscape of healthcare delivery, including policy changes and workforce pressures, understanding these associations can provide valuable insights into improving patient-centred care.

Methods

In Scotland, the Health and Care Experience (HACE) survey, conducted biennially since 2009–2010, gathers local and national insights into patient perspectives on health and care services. Mandated by the Scottish Government and administered by Public Health Scotland (PHS), HACE targets a random selection of patients registered with general medical practices. It seeks to capture personal experiences across several domains: general practices, out-of-hours healthcare services, caring responsibilities and support for care and daily living.

We conducted a secondary analysis of general practice (GP) patient experience survey question data from the HACE survey of patient satisfaction, which is administered biennially to a random sample of patients aged 17 years or older registered with general medical practices across Scotland, identified using the Community Health Index (CHI) database. Respondents are initially invited to complete the survey online, with a paper questionnaire sent as a reminder to non-responders. Sampling is conducted within GP practice lists, with the required sample size calculated separately for each practice to achieve a sufficiently reliable result; required sample sizes were capped at a maximum of 1000 per practice. For practices with very small numbers of eligible people, all were included to meet minimum sample size requirements. Comprehensive details of the survey have been published elsewhere.^7–13 The number of surveys sent out varied between years, ranging from 485 380 for 2009–2010 to 537 924 in 2021–2022. Survey results are reported as the percentage of respondents giving a positive response to each item, aggregated at practice level. The results of the surveys between 2009–2010, 2011–2012, 2013–2014 and 2021–2022 were obtained from the Scottish Government website,^{7 14–16} while the 2015–2016, 2017–2018 and 2019–2020 results were provided by PHS on request.¹⁷ Data relating to covariates including practice size, deprivation and rurality were also obtained from PHS.^18–20

The Scottish eightfold urban–rural classification (UR8) provides a standard definition of the rurality of areas and is updated every 2 years to incorporate the most recent ‘small area population estimates’ produced by National Records of Scotland and the Royal Mail postcode address file. Table 1 shows each UR8 classification and its code definitions. Further information regarding the methodology used for the population estimates for settlements can be found on the Scottish Government website.²¹ There were small numbers of practices in some UR8 classifications, so (UR3–5 and UR6–8) were merged for analyses. It should be noted that the UR8 classification does not account for specific considerations related to island communities and coastal areas. The classification provides a broad categorisation based on settlement size and driving time; however, it lacks nuance in relation to the unique transport situations and specific challenges faced by island residents, including potential transport isolation during winter. Coastal towns also face demographic and logistical challenges, such as higher healthcare demands from an ageing population.²²

Table 1.

Scottish eightfold urban–rural classification (UR8) and its definition for each code, settlement size and driving time⁴⁸

UR8 classification	Code definition	Settlement size and driving time
UR1	Large urban areas	Settlements of over 125 000 people
UR2	Other urban areas	Settlements of 10 000–125 000 people
UR3	Accessible small towns	Settlements of between 3000 and 10 000 people and within 30 min’ drive of a settlement of 10 000 or more
UR4	Remote small towns	Settlements of between 3000 and 10 000 people and with a drive time of between 30 and 60 min to a settlement of 10 000 or more
UR5	Very remote small towns	Settlements of between 3000 and 10 000 people and with a drive time of over 60 min to a settlement of 10 000 or more
UR6	Accessible rural areas	Settlements of less than 3000 people within 30 min’ drive to a settlement of 10 000 or more
UR7	Remote rural areas	Settlements of less than 3000 people and with a drive time of between 30 and 60 min to a settlement of 10 000 or more
UR8	Very remote rural areas	Settlements of less than 3000 people and with a drive time of over 60 min to a settlement of 10 000 or more

The Scottish Index of Multiple Deprivation (SIMD) is the national tool for classifying area deprivation across Scotland. It aggregates a wide array of indicators across several domains, including income, employment, health, education, geographic access, housing and crime, to construct a comprehensive picture of relative deprivation.²³ In this study, quintiles were based on the distribution of deprivation scores from the latest data release: quintile 1 represents the most deprived areas and quintile 5 the most affluent.

HACE survey responses were presented by general practice name, while the information regarding the UR8 classification and SIMD quintile of each practice was allocated based on the practice code. Each practice result from the HACE survey was matched with its UR8 classification and SIMD quintile based on its practice code. The UR8 classification allocation was based on the modal urban-rural category of the patients in each practice. UR8 classifications are updated biennially, so each year of survey results was matched to the corresponding list of UR8 classifications (supplementary table I). The SIMD quintiles were based on information relating to the last quarter of 2021–2022 data obtained from files requested and obtained from PHS—the modal value was used to utilise the quintile containing most patients.

Patient-centred care and continuity of care were assessed by questions in the HACE survey. Respondents were asked to indicate their level of agreement with the following statements: ‘I was listened to’, ‘I was given enough time’, ‘I was treated with compassion and understanding’, ‘I was given the opportunity to involve the people that matter to me’, ‘I understood the information I was given’, ‘I was in control of my treatment/care’ and ‘I knew the healthcare professional well’. The response alternatives were ‘very positive’, ‘positive’, ‘neutral’ and ‘negative’. Responses of ‘very positive’ and ‘positive’ were aggregated to give ‘percentage satisfied’.

The wording of the questions did not remain consistent for every survey prior to 2017–2018. To facilitate appropriate comparison across survey years, suitable alternative questions were identified from the previous surveys and then matched only when the meaning of the question was deemed the same and on agreement between the coauthors (online supplemental appendix I). Some questions from the recent survey could not be compared over time due to a lack of consistency. For those questions that were compared, we ensured they were used over at least three consecutive survey years. Supplementary graphs relating to UR8 classification, SIMD quintile and practice size, which only included data points with consistent wording across time were also created in a sensitivity analysis (online supplemental file 2).

Patients from 1008 GP practices participated in the 2009–2010 HACE survey. A proportion of these practices could not be identified on the 2021–2022 survey list. Some practices changed UR8 classification between these time periods, but practices that changed but remained within collapsed categories, such as from UR3 to UR4, were included in the analyses. However, practices that changed and did not remain within the collapsed categories, such as changing from UR3 to UR7, were excluded. In sensitivity analyses, practices with a change in registered list size of greater than or equal to 30% between the survey periods were excluded to account for the potential impact of practice mergers, closures or substantial organisational change. The findings were similar to those of the main analyses (online supplemental file 3).

Data analysis

A database was constructed to show practice name, code and population size, HACE survey result as a percentage of positive responses for each question analysed, alongside the UR8 classification and SIMD quintile. Averages were then calculated to give the mean percentage positive result for each survey year based on the UR8 classification and SIMD quintile. No sample weighting was applied, as excluded practices were likely to differ systematically from included practices in ways that make conventional weighting approaches inappropriate. Spearman correlation coefficients were used to examine associations between practice-level percentage positive responses to survey items. Statistical analyses were performed using the IBM Statistical Package for the Social Sciences V.29.0²⁴ for the correlation analyses and R statistical software (V.4.1.3), implemented via RStudio, for all other analyses.²⁵

Relative importance method

The relative importance method was employed to ascertain the predictive variables (UR8 classification, SIMD quintile and practice size quintile) that make the most substantial contribution to models of patient satisfaction within the 2021–2022 dataset. Relative importance analysis is a method used to assess the contribution of predictor variables in explaining the variance of the outcome variable in a regression model. It helps to determine which predictors are the most influential in predicting the outcome. This method complements multiple regression analysis by identifying the proportional influence each predictor variable exerts on the total variance explained in a regression model. This includes assessing each variable’s individual contribution and its synergistic effect with others. Relative importance is quantified as the average increment in R² when predictive variables are incorporated into a multivariable linear regression model, irrespective of their causal sequence. This approach allows for a further understanding of the predictive variables’ significance beyond the outputs of multiple regression analysis. Further information regarding the relative importance method has been published elsewhere.^26–28

Subsequently, to investigate the magnitude and direction of the important factors, the generalised linear mixed modelling method was employed, and both adjusted and unadjusted mean ratios were calculated. Model 1 examined the association between UR8 classification and patient satisfaction, unadjusted for other variables. Model 2 additionally adjusted for SIMD quintile. Model 3 was the fully adjusted model, incorporating UR8 classification, SIMD and practice population simultaneously.

Patient and public involvement

We did not directly include PPI in this study, but the database used in the study was entirely dependent on public perspectives on health services.

Results

Patients from 1008 general medical practices participated in the 2009–2010 HACE survey. Of these, 301 could not be identified on the 2021–2022 survey list, possibly due to closures or mergers. Of the remaining 707 practices existing in 2021–2022, 60 (8.5%) changed UR8 category between 2009–2010 and 2021–2022, and 647 practices remained consistent. Out of these, 15 practices did not specify their population size, therefore were also excluded. Data relating to 632 practices were, therefore, analysed in this study. While these practices can be considered to be representative of continuing Scottish general practice broadly, practices that closed or merged may have differed systematically from those that remained, and this should be considered when interpreting findings. HACE survey data were available for all 7 biennial survey years between 2009–2010 and 2021–2022, with response rates ranging from 16% to 38% across years, highest in the most remote and least deprived areas.

To assess the potential impact of practice mergers and major changes in practice list size over time, we conducted a sensitivity analysis excluding practices that experienced a ≥30% change in list size between 2010 and 2021. After applying this criterion, 590 practices remained. The results of this analysis were consistent with the main findings, with minimal differences in trends or overall conclusions, indicating that the main results were robust to potential artefacts introduced by practice mergers. Full details of this sensitivity analysis are provided in online supplemental file 3.

There were correlations between practice size, UR8 category and area deprivation as assessed by SIMD: see table 2.

Table 2.

Correlations between practice characteristics (3×3 table)

	UR8 classification	Mean final SIMD quintile	Total practice population
UR8 classification	1	0.269	−0.286
Mean final SIMD quintile	0.269	1	0.171
Total practice population	−0.286	0.171	1

SIMD, Scottish Index of Multiple Deprivation; UR, urban–rural classification.

Response rates to the HACE survey are relatively low,^7–13 see online supplemental appendix II, but compare well with response rates from other published surveys.²⁹ In the most recent HACE survey, the response rate ranged from 19% of people in large urban areas to 34% in remote rural areas, and from 15% of people in SIMD quintile 1 (most deprived) to 32% in quintile 5 (least deprived). These trends were seen for all years of the survey analysed (online supplemental appendix III and IV). Response rates did not differ substantially between patients in small and large practices (online supplemental appendix V). In every year of the survey analysed, women consistently had a higher response rate than men. Furthermore, the highest response rates were seen for those aged 65 years or more, and the lowest rates were seen in the youngest patients (see online supplemental appendices VI and VII).

It is important to note that the phrasing of survey questions was inconsistent prior to 2017–2018. To facilitate appropriate comparison across survey years, we matched questions from earlier surveys only when their meanings were comparable. This matching process was validated through consensus among the coauthors. A set of additional graphs relating to UR8 classification, SIMD quintile and practice size quintile, which only included data points with consistent wording across time, were also created as part of a sensitivity analysis, which did not alter our conclusions (online supplemental file 2).

The characteristics of practices in each collapsed UR8 category and SIMD quintile are given in table 3 and table 4, respectively.

Table 3.

Total number of practices in Scotland classified using Scottish eightfold urban–rural classification, mean, median, maximum and minimum practice patient list size in relation to 2021–2022 survey result data

UR8 classification	Number of practices (n)	Mean practice patient list size (n)	Median practice patient list size (n)	Maximum practice patient list size (n)	Minimum practice patient list size (n)
UR1	248	6475	6094	19 629	1248
UR2	188	8138	7690	25 037	1688
UR3–5	78	6948	6512	15 202	2023
UR6–8	118	3269	2926	9996	257

UR, urban–rural classification.

Table 4.

Total number of practices in Scotland classified using SIMD, mean, median, maximum and minimum practice patient list size in relation to 2021–2022 survey result data

SIMD quintile	Number of practices (n)	Mean practice patient list size (n)	Median practice patient list size (n)	Maximum practice patient list size (n)	Minimum practice patient list size (n)
1 (most deprived)	126	5212	4625	17 790	654
2	127	6525	6268	18 544	277
3	126	6610	6475	22 427	257
4	127	6348	6188	15 625	346
5 (least deprived)	126	6429	7284	25 037	675

SIMD, Scottish Index of Multiple Deprivation.

Secular time trends

Figure 1 shows the secular changes in responses to each of the HACE questions. Satisfaction levels captured by most questions remained broadly constant from 2009 to 2010 until 2019–2020, but there was a marked decline in reported satisfaction in the 2021–2022 survey.

Figure 1.

Time trends graph portraying mean percentage positive result for each satisfaction question analysed. Lines connect data points from repeated cross-sectional surveys conducted at different time points.

Online supplemental appendix VIII shows the correlations between percentage positive responses to survey items in the HACE 2021–2022 survey. In general, there are fairly strong correlations (r>0.53) between all the question items, although the item addressing personal continuity of care was less strongly correlated with other items (r=0.36–0.62).

Time trends according to UR8 category

Figure 2 shows the responses to individual survey items by collapsed UR8 category. Participants in the more rural UR6–8 areas consistently reported higher satisfaction levels over time.

Figure 2.

Data change over time based on the Scottish eightfold urban–rural classification for each satisfaction question analysed. Note difference in y-axis scales between graphs. Lines connect data points from repeated cross-sectional surveys conducted at different time points.

Time trends according to SIMD quintile

Figure 3 shows trends in patient satisfaction according to SIMD quintile. In general, participants registered with practices in the more affluent areas expressed moderately higher levels of satisfaction, with the exception of responses to ‘I knew the healthcare professional well’ and ‘I was given the opportunity to involve the people that matter to me’. For these items, patients in the more deprived areas provided substantially more positive feedback, with marked contrasts between the most deprived and other SIMD quintiles.

Figure 3.

Data change over time based on quintiles relating to the Scottish Index of Multiple Deprivation (SIMD) for each satisfaction question analysed. Difference in y-axis scales between graphs. Lines connect data points from repeated cross-sectional surveys conducted at different time points.

Time trends according to practice size

Figure 4 shows that patients registered with smaller practices expressed higher levels of satisfaction across various care and experience dimensions, most markedly in response to the question addressing continuity of care.

Figure 4.

Change in satisfaction responses over time based on quintiles relating to practice population size. Difference in y-axis scales between graphs. Lines connect data from repeated cross-sectional surveys conducted at different time points.

Relative importance analysis

Relative importance analyses illustrate the relative contributions of UR8 classification, SIMD quintile and practice size to satisfaction expressed in relation to the various survey items in the 2021–2022 dataset: see figure 5.

Figure 5.

Relative importance analysis. The relative importance percentage R² of the predictive variables assessed including a) UR8 collapsed classification b) SIMD quintile and c) practice size, assessed for each of the survey items patient satisfaction levels for the latest dataset analysed (2021–2022). It should be noted that this only includes the variation for the three factors mentioned, and not overall variation. Estimates are adjusted to total 100%. SIMD, Scottish Index of Multiple Deprivation; UR, urban–rural classification.

To further investigate the magnitude and direction of associations between the predictor variables and patient satisfaction, regression models were constructed for each survey item. Full results including unadjusted and adjusted mean ratios are presented in supplementary table II.

Discussion

Patient satisfaction in these surveys underwent a marked decline after 2019–20, coinciding with the COVID-19 pandemic.³⁰ The pandemic exacerbated challenges in healthcare delivery, leading to increased waiting times, reduced face-to-face consultation and a rapid shift to telehealth services.^{31 32} The pandemic also negatively impacted the mental health of healthcare workers, which could have subsequently affected patient care.³³ The 2021–2022 survey revealed a marked decline across all questions compared with previous survey years, with particular satisfaction reductions noted in the areas of ‘I was in control of my treatment/care’, ‘I was given the opportunity to involve the people that matter to me’ and ‘I knew the healthcare professional well’. This trend could reflect the impact of COVID-19 restrictions, which likely contributed to patients’ inability to bring relatives to appointments and the cancellation of elective surgeries. Consequently, there was an increased need for patients to attend GP appointments more frequently. Additionally, the shift towards virtual appointments rather than face-to-face interactions, combined with the general lack of control experienced in all areas of life due to lockdown measures, may have further exacerbated patients’ feelings of disempowerment and isolation.

Practice list size, socioeconomic status and rurality appear to be strongly intercorrelated, but each of these characteristics appears to make an independent contribution to patient satisfaction with care from general practice as assessed by HACE survey responses. Patients living in more rural areas (UR6–8), those from the least deprived datazones (SIMD quintile 4 and quintile 5) and those registered with smaller practices (in the lowest practice size quintile) generally reported the highest levels of satisfaction. Conversely, individuals living in suburban and small towns (UR2 and UR3–5), those registered with practices covering more deprived areas (quintile 1 and quintile 2) and patients of larger practices (in the highest practice size quintile) expressed more dissatisfaction. These findings align with literature, indicating that practice size, rurality and socioeconomic status significantly impact patient experiences and satisfaction levels within general practice services^{34 35} and that there are significant challenges faced by practices in deprived areas, including higher disease burdens and resource constraints.

One exception to this pattern was in responses to the HACE question ‘I knew the healthcare professional well’, where patients in practices in the most deprived SIMD quintile consistently reported the highest levels of satisfaction with this aspect of continuity of care. This may reflect greater levels of comorbidity in areas of deprivation,³⁶ leading patients to choose familiar clinicians to manage their care. This observation aligns with previous findings, which noted that dissatisfaction with healthcare services was often tied to a lack of familiarity with healthcare professionals,³⁷ suggesting that continuity of care with known clinicians can significantly impact patient satisfaction. Further studies have assessed the impact of socioeconomic and healthcare provision factors on patient satisfaction, indicating the significant role that familiarity with healthcare professionals plays in patient contentment.³⁸

The paradox of improving access to care while maintaining personal continuity is a core challenge in primary care. Recent evidence has suggested that as practices reactively implement rules to enhance access, continuity diminishes, leading to a rise in unmet patient needs.³⁹ Additionally, as the number of full-time equivalent GPs decreases, there is a growing reliance on other healthcare professionals, such as advanced nurse practitioners and pharmacists. While the contribution of these professionals is important, there is a risk of increasingly fragmented care. Consequently, fewer GPs, increasing demand, longer waiting times and the growing use of allied health professionals to manage the workload results in more transactional and episodic care.

The difference in level of patient satisfaction between rural and urban patients was amplified following the COVID-19 pandemic. This may reflect greater resilience and adaptability among rural practices in these circumstances. The underlying greater level of satisfaction in rural areas may reflect a greater range of health services being offered close to patients’ homes, or stronger patient-practice relationships in the context of a more limited choice of providers.

The relative importance analysis of the most recent survey data suggests the dominant role of socioeconomic status and practice size in influencing patient satisfaction across most survey items, with, to a lesser extent, rurality also being significantly associated with satisfaction levels.

Strengths and limitations of study

The methods employed in the government-mandated HACE surveys have been largely consistent over time, and general practice in Scotland is publicly funded and free at the point of use to all citizens. Patients registered with all practices in Scotland were surveyed, although data from practices that merged, closed or underwent significant demographic change in patient populations were excluded from the analysis of temporal trends. It is possible that survey results from patients registered with closed or merged practices may have differed substantially from those from the majority of respondents. While acknowledging this limitation, it is likely that the populations registered with the practices that were excluded may have become increasingly dissatisfied over time due to practice closures or mergers. The practices which were excluded would likely have differed systematically from those that remained in the study and it is unlikely that weightings would make our findings more robust. We were also not able to capture specific details about the most rural practices and settings, another issue worthy of future investigation. As only aggregated practice-level data were available, it was not possible to account for potential clustering of patients within practices through multilevel modelling, which represents a further limitation of this analysis.

This study analysed aggregate practice-level data, with satisfaction expressed as the mean percentage of positive responses across all respondents within each practice. Findings, therefore, describe associations at the practice level and cannot be directly extrapolated to individual patients. Standard errors and precision estimates reflect between-practice variability across the 632 included practices, rather than within-practice patient-level variation, which is not captured in this analysis. Results should, therefore, be interpreted as reflecting characteristics of practices and their registered populations rather than the experiences of individual patients.

Response rates to the HACE survey are relatively low and they varied with demographic factors significantly, with lower participation observed in urban and deprived areas. The under-representation of younger individuals and males may reflect their more limited usage of primary care services, and this should be taken into account in interpretation of the findings.

Responses to the HACE survey items are generally highly intercorrelated, suggesting that they reflect a broad aspect of satisfaction with general practice. The analysis of correlations between the percentage positive responses to questions in the HACE surveys revealed robust correlations among the majority of the question items. However, the item pertaining to personal continuity of care demonstrated weaker correlations with the other items. Some items (eg, 'I was given the opportunity to involve the people that matter to me’) are nevertheless unlikely to be applicable to all patient encounters and their interpretation may, therefore, be difficult.

Despite the availability of data spanning multiple timepoints, variations in the HACE survey questions over time necessitated the comparison of similar, rather than identical, questions prior to 2017–2018, which might have compromised the validity of some of the results. This may explain reductions in satisfaction levels in some years such as 2013–2014 and 2015–2016 for survey items such as ‘I understood the information I was given’ and ‘I was in control of my treatment/care’.

Further HACE survey data continue to be released, offering the potential for future research into trends and underlying mechanisms.

Implications

Satisfaction with primary care services to some extent reflects quality of care⁴⁰ and is likely to be associated with favourable health outcomes.^41–46 Interventions aimed at improving patient satisfaction must consider the complex interplay between socioeconomic factors and organisational aspects of care delivery. Disparities in patient satisfaction could be reduced by improving accessibility of services⁴⁷ in more deprived and urban areas and by measures to improve personal continuity of care. There is a need for more research on the impact of different primary care delivery models on patient satisfaction, including a focus on the use of telehealth services and its implications for different sociodemographic groups. Specifically in regard to national surveys, attention needs to be paid to wording, both in terms of consistency over time and the need to capture patient experiences accurately. Conducting rigorous science is challenging when it is not solely guided by scientific principles. Continued study of these data could potentially serve as a valuable indicator of the impact on patients of change and instability in primary care. Nevertheless, in more recently published HACE survey data, the wording of survey questions has been changed to an extent that creates difficulties in comparing results to previous years.

Data availability statement

Data are available upon reasonable request. The data used in this study are derived from the Scottish Health and Care Experience (HACE) survey, administered by Public Health Scotland (PHS) on behalf of the Scottish Government. Some survey data are publicly available through Public Health Scotland and the Scottish Government website. Additional linked datasets used in this study (Scottish Index of Multiple Deprivation, urban-rural classification, and practice population size data) are also available through the Scottish Government. Some data were obtained upon request from Public Health Scotland.

Ethics statements

Patient consent for publication

Not applicable.

Ethics approval

Ethics committee approval was not required as all data were available in the public domain, either online or by direct request to the Scottish Government Care Experience Survey Team. This was confirmed as appropriate by the head of department at the time this research commenced. Contact details: Professor Philip Wilson, Professor Emeritus, The Centre for Rural Health, University of Aberdeen. Email: p.wilson@abdn.ac.uk