Abstract
Background Professionals in general practice (family medicine) experience more violence and abuse by patients and the public than in other healthcare settings. There is limited research on such experiences among non-clinical staff, and no direct comparisons between staff groups in general practice.
Aim To explore the following: the extent of violence and abuse from patients or the public towards general practice staff between 2020 and 2023; staff correlates and environmental correlates for violence and abuse; and potential impacts of violence and abuse regarding staff feeling of safety and support at work.
Design & setting An online survey of general practice staff was conducted across England between 11 July 2023 and 30 November 2023.
Method The survey questions covered demographics, physical violence and threats, verbal abuse, harassment, and inappropriate sexual behaviours experienced or witnessed between 2020 and 2023. It asked whether participants felt safe and supported at work. Mann–Whitney U tests were used to compare binary outcomes and Kruskal–Wallis tests were used for categorical outcomes.
Results Participants (n = 1152, 44.0% clinical, 56.0% non-clinical) were aged 21–75 years (mean = 47.3 years, standard deviation [SD] = 11.1). Overall, 93.7% reported experiencing any form of violence or abuse, with 92.3% reporting verbal abuse, 47.7% reporting physical violence or threats, 60.5% reporting feeling harassed, and 23.7% reporting inappropriate sexual behaviours. Additionally, 22.3% of staff felt unsafe but only 57.1% felt supported at work. Non-clinical, younger, or less experienced staff and those in urban and deprived areas experience more violence and abuse. Those experiencing it more frequently felt less safe and supported.
Conclusions Violence and abuse from patients and the public towards general practice staff may be prevalent and increased since the COVID-19 pandemic. Those at greater risk require more organisational support.
How this fits in
This is the first nationwide study in England to include direct responses from both clinical and non-clinical staff, and the first workplace violence study applying poly-victimisation by analysing correlations between levels of exposure and consequences. It found approximately one-fifth of the participants felt unsafe but only 57.1% felt supported for their experience of violence and abuse. The findings indicate the need for prevention strategies and staff support at organisational and system levels.
Introduction
Levels of violence and abuse towards staff from patients in general practice (family medicine) may be greater than in hospitals, with a UK survey suggesting more than 50% of GPs, compared with 31% of hospital doctors, had experienced abuse from patients in the previous month.1 Such experiences impact staff health,2 intention to leave, and service delivery.3 However, general practice receives less attention than other healthcare settings in this regard, both in research and from organisational support and interventions;4 for example, in the UK, general practice is not included in the annual NHS Staff Survey.5
Research on violence and abuse by patients and the public towards general practice staff (referred to as ‘violence and abuse’ from this point) mostly focuses on GPs,4 although evidence suggests that care navigators (formerly termed ‘receptionists’) and non-clinical staff experience more violence and abuse than clinicians.1,6 Understanding different experiences across staff groups is key to target support appropriately. A recent scoping review solely focused on patient-initiated violence4 found those who are younger7,8 or with fewer years in general practice9,10 experienced more violence. Professionals working in practices in more socioeconomically deprived areas experienced more violence,11 while evidence on rurality was inconsistent.12,13
It is important to establish both prevalence estimates of violence and abuse and frequency levels of individual experiences among participants because multiple exposures to any type may lead to greater health issues and reduced wellbeing.14 While prevalence estimates show the extent of the problem among participants, understanding exposure levels helps inform the types of support staff may need. Evidence suggests reasons for violence and abuse may include perceived unmet patient needs, including increased waiting times15,16 and general frustration with health services.17,18 As waiting times for patients to be seen across healthcare settings increased dramatically since the start of the COVID-19 pandemic in 2020,19 consideration of whether violence and abuse towards staff has also increased over this time period is important.
We therefore aimed to explore:
the extent and type of direct or witnessed violence and abuse experienced by general practice staff from patients or the public between 2020 and 2023;
the association between staff correlates (age, the number of years working in general practice, and job roles) and the frequency of violence and abuse experienced by general practice staff from patients or the public;
the association between environmental correlates (rurality, deprivation, and geographical regions) and the frequency of violence and abuse experienced by general practice staff from patients or the public; and
potential impacts of violence and abuse from patients and the public on general practice staff regarding feeling safe and supported at work.
Method
Design
An online survey of general practice staff in England was conducted from 11 July 2023 to 30 November 2023 as part one of a mixed-methods study.20 It was promoted via National Institute for Health and Care Research (NIHR) Clinical Research Networks (CRNs; now Research Delivery Networks) across England, the national Society for Academic Primary Care e-newsletter, the social media site X (formerly Twitter), and locally via the University of Nottingham Primary Care Education Unit distribution lists.
The survey contained 19 questions, adapted from relevant items in the NHS Staff Survey and piloted by two GPs, a patient and public involvement and engagement (PPIE) adviser, and a practice manager. Participants were asked about the frequency (never, 1–2 times, 3–5 times, 6–10 times, or >10 times) of experiencing or witnessing any violence or abuse, including physical abuse and threats, verbal abuse, harassment, and inappropriate sexual behaviours, in the preceding 3 years. Participants were asked whether such behaviours had happened more, less, or just as often over this period (that is, whether they had increased in frequency, decreased, or stayed the same) and whether measures within their practice made them feel ‘safe’ or ‘supported’ (possible answers: strongly agree, agree, neutral, disagree, or strongly disagree). Participants were asked whether they were having strong reactions such as flashbacks, panic attacks, severe depression, or thoughts about hurting themselves (possible answers: yes, maybe, no). Demographic questions included age and gender (with a ‘prefer not to say’ option), years worked in general practice, and the postcode of the GP practice currently or most recently worked in (if not currently working in general practice). The qualitative analysis is reported elsewhere.21
Participants
All staff aged ≥18 years who had started working in general practice before March 2020 and worked through the COVID-19 pandemic were eligible to participate.
Procedure and ethical considerations
The survey was hosted on the secure JISC platform. Participant information was provided after a screening question on whether they had worked in a GP surgery before and during the COVID-19 pandemic. Active rather than implied consent was requested before the survey questions. If participants withdrew by closing the browser, prior answers were not saved. The survey ended with a debriefing page with information on support organisations.
Data management
Survey data were exported to Microsoft Excel (version 2308), with regions, rurality, and Index of Multiple Deprivation 2019 (IMD) subsequently added, all based on practice postcodes. Practice rurality was extracted from the UK 2011 Rural Urban Classification,22 ranging from 1 = most urban (major conurbation) to 8 = most rural (hamlets and isolated dwellings). IMD was in percentiles, with 1 = the most deprived and 99 = the least deprived. Practices were grouped into four geographical regions: South West, South East or South Midlands, the Midlands, and North. This grouping ensured sizes suitable for comparisons. Job roles were described in relevant groups, then also categorised into clinical and non-clinical groups for some analyses.
Data analysis
Demographic data were reported using means and standard deviations (SD) for normally distributed data or median and interquartile range (IQR) for non-normally distributed data. Numbers and percentages of reported frequency of violence and abuse experienced or witnessed were calculated. A binary variable for occurrence of each type of abuse was also produced: ‘No’ (Never) versus ‘Yes’ (1–2 times, 3–5 times, 6–10 times, more than 10 times combined). The occurrences of violence and abuse were compared across clinical and non-clinical groups using χ2 tests. To identify levels of feeling unsafe or unsupported, responses for these questions were dichotomised into disagree or strongly disagree (unsafe or unsupported) versus neutral, agree, or strongly agree (not unsafe or unsupported). Feeling unsafe or unsupported was then compared across clinical and non-clinical groups using χ2 tests.
An overall frequency level of violence experienced was calculated by scoring individual’s responses to each violence reporting question as follows: Never = 0, 1–2 times = 1, 3–5 times = 2, 6–10 times = 3, >10 times = 4. When examining correlation, skewness and kurtosis were calculated for all variables, showing that frequency of violence and abuse was not normally distributed. Spearman’s correlation coefficients were used to explore associations between the frequency levels of violence and abuse, and rurality, deprivation, age, and years in general practice.
Owing to non-normally distributed levels of violence and abuse, Mann–Whitney U tests were used when comparing binary outcomes and Kruskal–Wallis tests were used for categorical outcomes. To compare occurrence of abuse (yes versus no) experienced by clinical versus non-clinical staff, and then by individual staff groups, logistic regression was used to calculate unadjusted and adjusted odds ratios (aORs). When examined by individual staff group, GPs (the largest group of responders) were used as the reference group. Gender, age, years of experience, rurality, and deprivation were adjusted for. These covariates were selected a priori based on existing literature showing associations with patient violence exposure in general practice.4 Data were analysed using Stata (version 18).
Results
We had 1226 responses. Four were from outside England, two had invalid postcodes, and 68 did not meet our eligibility criteria and were therefore excluded. This left 1152 valid responses. Participants (total: n = 1152, with n = 507 [44.0%] clinical, and n = 645 [56.0%] non-clinical staff) were aged 21–75 years (mean = 47.3 years, SD = 11.1), with 83.7% self-identified as women, 15.5% as men, and 0.8% other or prefer not to say. GPs formed the largest group (26.0%), followed by practice managers (18.7%), and care navigators (18.1%). Most participants (86.4%) worked in urban areas. Participant characteristics are in Table 1.
Direct and indirect experience of violence and abuse
Odds ratios for direct and indirect experience of violence and abuse between non-clinical and clinical staff are shown in Table 2. Most participants (93.7%) reported experiencing any form of violence and abuse, with 92.3% reporting verbal violence and abuse (89.9% clinical, 94.1% non-clinical); 47.7% physical violence and abuse or physical threats (35.1% clinical, 57.5% non-clinical), 60.5% (45.6% clinical, 72.2% non-clinical) harassment (repeated intrusions); and 23.7% inappropriate sexual behaviours (29.0% clinical, 19.5% non-clinical). Furthermore, 96.3% of participants reported witnessing violence and abuse towards their colleagues, with 95.9% having witnessed verbal abuse, 57.1% physical violence or threats, and 26.0% inappropriate sexual behaviours. Most participants (62.2%) felt violence and abuse increased across this period, with more non-clinical than clinical staff sharing this observation (P = 0.04).
Staff roles
Among all participants, 92.1% of clinical staff and 94.9% of non-clinical staff reported experiencing any form of violence and abuse in 2020–2023 (Table 2). Care navigators (formerly ‘receptionists’) were the most likely to report (99.5%), while administrators were the least likely (88.3%) (Supplementary Table S1). Non-clinical staff experienced all types of violence and abuse more frequently except inappropriate sexual behaviour (Tables 2 and Table 3). After adjusting for gender, age, years worked in practice, rurality and deprivation, non-clinical staff had 2–3.5 times higher odds (Table 2) of experiencing violence and abuse than clinical staff (for example, feeling harassed by patients or the public, aOR 3.54, 95% confidence interval [CI] = 2.70 to 4.66, P<0.0001), other than for inappropriate sexual behaviours which non-clinicians had 56% lower odds of experiencing than clinicians (aOR = 0.44, 95% CI = 0.33 to 0.60, P = 0.001).
After adjusting for gender, age, years worked in practice, rurality and deprivation, care navigators had the highest odds of experiencing verbal abuse, over 14 times that of GPs (aOR 14.54, 95% CI = 1.90 to 111.28). Practice managers had over four times the odds of experiencing physical abuse or threats compared with GPs (aOR 4.21, 95% CI = 2.22 to 5.08). Practice managers and care navigators were over four times as likely as GPs to experience harassment (aOR 4.59, 95% CI = 2.95 to 7.15 for practice managers; aOR 4.49, 95% CI = 2.80 to 7.20 for care navigators). Nurses and healthcare assistants had roughly twice the odds of experiencing sexual abuse compared with GPs (aOR 1.95, 95% CI = 1.19 to 3.20). Full adjusted models with all covariates are presented in Supplementary Table S1.
Staff age and experience
Younger age was correlated with more frequent violence and abuse, shown by the negative correlation between these two factors (r = −0.115, P<0.001, Table 4). Similar significant patterns were shown for all subtypes of violence and abuse, except feeling harassed (r = −0.012, P = 0.68). Shorter time worked in primary care was also correlated with higher levels of violence and abuse (r = −0.132, P<0.001).
Rurality and deprivation
Staff from practices in more urban and more deprived areas reported higher levels of violence and abuse (r = −0.136, P<0.001; r = −0.145, P<0.001, respectively, Table 4). Verbal abuse was reported more in more deprived areas (r = −0.119, P<0.001), but showed no correlation with rurality (r = −0.053, P = 0.07). Inappropriate sexual behaviour was reported more in more urban areas (r = −0.062, P = 0.04), but not correlated with deprivation (r = −0.004, P = 0.9).
Geographical regions
No significant association was found between the frequency levels of violence and abuse and geographical region.
Possible impacts of violence and abuse
Overall, 22.3% of staff disagreed or strongly disagreed with the statement that they felt safe at work, while 57.1% of staff reported feeling supported at work (Table 5) . There was no significant difference in feeling unsafe or unsupported at work between clinical and non-clinical staff (Table 2). Increasing frequency of violence or abuse was correlated with feeling less safe (r = −0.3, P<0.001) and feeling less supported (r = −0.210; P<0.001). Meanwhile, 4.0% of participants reported experiencing symptoms such as flashbacks, panic attacks, severe depression, or thoughts about hurting themselves at the time of survey completion.
Discussion
Summary
Among our participants, experiences of violence and abuse were widespread, with 62.2% reporting increases during the 2020–2023 period. We identified substantial differences in experiences between staff roles. Those younger or with less work experience in general practice, and urban practices or practices in more deprived areas experienced more frequent violence and abuse. Non-clinical staff had twice the odds of encountering violence and abuse versus clinical staff. Care navigators faced the highest odds of verbal abuse. Practice managers and care navigators experienced the most physical abuse, threats, and harassment. Inappropriate sexual behaviour, although less common than other types of violence and abuse, was experienced more by clinicians, especially nurses and healthcare assistants. Overall, 22.3% of participants reported not feeling safe and only 57.1% felt supported at work. More frequent violence correlated with staff feeling less safe and supported.
Strengths and limitations
This is the first nationwide study in England including direct responses from both clinical and non-clinical staff, and the first workplace violence study applying poly-victimisation by analysing correlations between levels of exposure and consequences. Care navigators were well represented in the study sample (18.1% of participants), and the sample size and spread across different staff groups allow between-group comparisons. However, like many existing staff surveys on workplace violence in primary care, convenience sampling was used without a pre-determined sampling frame. It is therefore not possible to know the response rate or systematic differences between responders and non-responders, making it difficult to draw firm conclusions about actual prevalence. Nevertheless, the high responses within a short recruitment period from an overworked workforce may highlight the importance general practice staff accorded to this issue. Furthermore, the questionnaire used was not validated. However, owing to time and workload pressures on general practice staff, questionnaires validated on hospitals (for example, the Perception of Prevalence of Aggression Scale questionnaire)23 may be too long to attract responses. One question asked about actual or threatened physical violence together, preventing determination of the proportion experiencing actual versus threatened physical violence. While the 3-year recall period may increase recall bias in frequency estimates, it was deliberately selected to capture experiences during COVID-19, when systems faced unprecedented workforce pressures.
Comparison with existing literature
While our sampling methods limit prevalence comparisons, including all main staff groups enabled meaningful between-group comparisons that previous surveys focusing on single roles could not achieve. Unlike prior studies combining multiple settings and perpetrator types without distinction, or focusing on single regions, our survey covered all of England. We examined a 3-year period including the COVID-19 pandemic rather than the typical 1-year or career-long timeframes. This may explain why our observed rates of violence and abuse exceed reports from primary care staff in other countries during a similar period (18.3% in China),24 yet still surpass career-long rates previously reported in the UK (70.3% in 2013)25 and those from hospital staff (23.9%–87.5%)26 and pharmacists (45% [95% CI = 30% to 60%]).27
Nevertheless, our findings that non-clinical staff had greater odds of experiencing violence and abuse than clinical staff aligns with the 2021 British Medical Association (BMA) survey1 and a 2018 Spanish study.28 Similarly, staff with less experience9,13 or younger staff29,30 were more likely to encounter violence and abuse. Fear and apprehension at work31,32 have also been demonstrated previously.
We identified new findings not previously reported in studies on patient violence. Higher levels of experienced violence and abuse corresponded with staff feeling less safe and supported. This cumulative impact of multiple exposures (poly-victimisation) has not been discussed in workplace violence research. Only 57.1% of participants felt supported regarding their experience of violence and abuse. Despite different levels of violence exposure, feelings of being unsafe and not supported are shared across staff groups. Our findings also show that practices in urban and deprived areas were more exposed to violence and abuse whereas previous findings on rurality and abuse were inconsistent.12,13 The experiences did not appear to vary across English regions, a finding not previously demonstrated.
Implications for research and practice
Less than 60% of our responders felt supported at work. Interventions to prevent violence and abuse and management support for those who experience violence and abuse are needed. To protect and support those most at risk, some interventions may need to be targeted at practices in urban and deprived areas or at the staff groups most exposed to violence and abuse. It is important to provide more support to those who experience more violence and abuse. Given management staff did not feel significantly safer or better supported than other staff, certain interventions may need to be at a systems level, such as at a general practice, primary care network (groups of practices working with other health and social care providers), or integrated care system (local partnership for specific geographical areas) level.
Front-desk care navigators are the first point of contact and were the most exposed to violence and abuse. To the best of our knowledge, the reasons for younger and less experienced healthcare staff experiencing more violence and abuse have never been directly studied and have only been speculated about in post-research interpretations. However, research in the service industry found that customers’ biased views on age and inexperience contribute to increased aggression towards younger workers.33 Therefore, qualitative research into the antecedents for and responses to violence and abuse at individual and system levels would enrich understanding of this issue and its context, and help develop tailored prevention and support interventions for general practice.
Notes
Funding
This study/project is funded by the National Institute for Health and Care Research (NIHR) School for Primary Care Research (project reference 614). The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care.
Ethical approval
The study received approval from Health Research Authority (23/HRA/2515) and favourable opinion from the University of Nottingham Faculty of Medicine and Health Sciences Research Ethics Committee (240-0423).
Provenance
Freely submitted; externally peer reviewed.
Data
The dataset relied on in this article is available from the corresponding author on reasonable request.
Acknowledgements
We thank Luke Sawyers for his help in setting up the survey on JISC and advertising the study. Most importantly, we thank everyone who gave up their time to respond to our survey and those who helped distribute the survey.
Competing interests
The authors declare that no competing interests exist.
- Received June 27, 2025.
- Revision received October 21, 2025.
- Accepted November 12, 2025.
- Copyright © 2026, The Authors
This article is Open Access: CC BY license (https://creativecommons.org/licenses/by/4.0/)
References
LinkedInJump to section
More in this TOC Section
Related Articles
Cited By...
