Qualitative analysis

Included in the survey were two open questions; ‘What is the greatest problem within general practice at the current time’ and ‘What intervention would help general practice the most?’. The free-text comments were imported into NVivo V.11 and analysed with a thematic approach.17 Following a period of familiarisation, TS and TH developed an initial coding framework by coding a subset of 100 of the comments independently. This was reviewed by the full research team, and the agreed coding framework was then applied to the free-test data. The higher order categories were linked to the quantitative analysis in order to supplement and expand the interpretation of the data, and illustrative quotes were selected.

Quantitative analysis

Basic descriptive statistics were used to characterise the survey population and compare it to Health Education England data.5 Binary logistic regression analysis was employed to identify predictors of GPs’ intentions to retire within 5 years using a range of covariates, which were entered into the model simultaneously; gender, age, hours of work, role, length of service, job satisfaction.

Participants were provided with an information sheet outlining the study and were informed that completion of the online questionnaire would be taken as consent to participate.

Patient and public involvement

Patient and public involvement was not included in this study. The research question, although important to patients and the public, was focused on professional and health service priorities and experiences.

Participants

The survey was distributed by email to the 1697 GPs listed as working in Wessex, leading to 929 (54.7%) respondents. Of these, 509 (54.8%) were female, the modal age was 45–55 years (n=253, 32.9%) and most had been trained in the UK (93.0%). When compared with National Health Service (NHS) demographic data for all GPs in Wessex, there was no difference in gender balance, but there was a difference in age distribution, with our survey having an over-representation of older GPs (28.4% aged greater than 55 years compared with 20.1% in the NHS data; χ²=20.6, p<0.001).

When compared with the 2014 survey respondents, the current survey included more older GPs (28.4% aged greater than 55 years, compared with 23.7% previously) and more who were working in non-principal roles (41.5% compare to 22.6% previously), see table 1.

Nearly half of the respondents had spent over 20 years in general practice; one-third reported working over 41 hours per week; and nearly two-thirds reported having at least one additional employed role in addition to their NHS GP clinical responsibilities.

The two open questions had high completion rates (n=807, 86.9%; n=819, 88.2%, respectively), and the answers together provided a dataset of 29 679 free-text words; individual responses ranging from 1 to 340 words (mean=18).

Changes in work volume, intensity and morale

Respondents reported working an average of 29.6 hours a week (range 1–66) of which an average of 20.1 hours (range 2–59) were in direct contact with patients. As shown in table 2, the number of hours worked varied by employment status, with almost half of GP principals working 41 hours or more per week, while the most salaried GPs worked fewer than 30 hours per week and the majority of locum GPs worked fewer than 20 hours.

Comparing current workload with 2 years previously, 51.0% (470) reported working longer hours with almost all (94.4%; 423) giving increased workload as the predominant reason; 26.6% had reduced their hours of work, with most (72.3%; 172) stating this was due to increasing intensity of workload and for many (29.8%; 71) it was related to stress and mental health. This contrasts with the intentions stated in the previous survey where 21.5% of GPs wished to reduce their hours worked and only 2.7% wished to increase.

Morale was reported as having reduced over the past 2 years for 59.4% (510) of respondents and increased for 14.1% (121). In total, 28.9% (247) now reported having positive morale and 42.7% (365) negative morale.

Intention to leave GP

When asked to think about their career plans compared with 2 years ago, 409 (48.5%) said they had brought forward their plans to leave GP, with just 47 (5.6%) planning to remain longer. Intention to leave/retire in the next 2 years has increased from 13% in 2014 to 18% (p=0.02), while 63.9% reported an intention to continue working for at least the next 5 years in 2014 compared with only 48.5% in 2017 (p<0.0001) (see table 3).

Binary logistic regression of GPs planning to retire or leave GP (see online supplementary file 2) identified those aged between 55–59 years and 60–64 years were much more likely to express an intention to leave, when compared with those aged 25–34 (OR 7.98; 95% CI 2.6 to 24.1; p<0.001, OR 7.1; 95% CI 1.7 to 30.0; p<0.01, respectively). Likewise, those who have served 20–29 years in GP were more likely to express an intention to leave when compared with those with less than 5 years of service (OR 3.3; 95% CI 1.3 to 8.3; p<0.05). Lower job satisfaction over the past 2 years was also significantly associated with intention to leave (OR 4.2; 95% CI 2.3 to 7.6; p<0.001).

A further regression, controlling for age and gender (see online supplementary file 2), showed that there was a modest association between having reduced working hours over the past 2 years and an intention to leave GP completely (OR 1.595; 95% CI 1.062 to 2.397, p<0.05).

Respondents were asked to rate on a Likert scale (1=not important, 5=very important) factors that might be contributing to their intention to leave GP (table 4). Intensity of workload had the greatest influence (mean=4.4) followed closely by volume of workload (mean=4.3) and too much time spent on unimportant tasks (mean=4.0). Lack of patient contact, potential introduction of a 7-day working week and reduced job satisfaction also scored a mean >3. Personal factors of note were age (mean=3.5), medical indemnity payments (3.4) and increased risk of litigation (3.0). They were also asked to rate factors that might help retain them in GP (table 4), again confirming the importance of addressing the volume and intensity of workload.

Current challenges to general practice

Analysis of the responses to the open question ‘What is the greatest problem within general practice at the current time?’ yielded five key themes: increasing demands, expectations and complexity of patients; workload; GP recruitment and retention; inadequate funding; and bureaucratic and administrative burden.

Suggestions for improving general practice

Responses to the open question ‘What intervention would help general practice the most?’ revealed eight themes. The number of respondents with answers that included each theme is shown in table 5.

Greater funding

Increasing funding for general practice was viewed as the most important requirement. Many participants felt that other problems, such as the lack of GPs and meeting patients’ expectations, could only be tackled with greater funding.

It’s all money. More money to pay extra GPs and pay GPs more to attract good doctors and retain the drs once trained (ID 220)

Positivity towards, awareness and experience of national workforce initiatives

Respondents were asked to rate whether they thought about the nationally led initiatives that had been recently introduced to address workforce issues in general practice, specifically whether the initiatives would have a positive, negative or no impact. A net rating was calculated by subtracting the percentage who rated the initiative as negative from the percentage who rated it as a positive. As shown in table 6, investment in practice nursing, improved access to specialist advice, investment in technology and expansion of the GP workforce were viewed most favourably, with greater than 75% net rating scores. Conversely, the wider use of physician associates (PAs), local sustainability and transformation plans (STPs), and video and e-consultations were rated negatively. There were mixed levels of awareness about the various national workforce initiatives being implemented, and with few exceptions (closing working with specialists, increased use of pharmacists, involvement in GP federations) the majority of respondents lacked direct experience of them (table 6).

Age emerged as having a small, but statistically significant correlation with attitude towards several individual initiatives (table 7). For example, younger GPs were more likely to be positive about federations, but were less positive in their views of PA. However, the attitudes towards most of the initiatives were very similar across all age groups.

Having had experience of an initiative was associated with a more positive attitude score towards it. The differences in mean scores were modest, but for seven of the initiatives, the difference was statistically significant (table 8).

One hundred and ninety GPs gave free-text comments to explain their views. The most widely stated theme was that there were too many initiatives, of which little or no benefit was being seen on the ground.

There are too many initiatives. GPs just need to be left alone to get on with the job with adequate funding. These initiatives cost money which comes out of GP budgets ID 925

Many of these ideas are great on paper but little evidence of impact at the coalface ID 826

Many felt that these initiatives were a distraction from the need for significant investment in general practice and tackling key issues affecting the workforce.

The only thing that will make any real improvement in care is investment in proper well-trained GPs continuing to be the centre of patient care in primary care alongside practice nurses with a proper career structure and practice pharmacists. All the other initiatives are just tinkering at the edges - smokescreens to try to take the heat off the central issue of lack of investment in General Practitioners ID 688

An additional theme suggested that some initiatives could be further undermining GP morale.

I object to the term ’resilience' and any resources invested into it. We should be focusing all our intentions on making the job better rather than coaching GPs to be more robust against the stress. The very term makes it sound to me like it is somehow the GPs fault in the first place for not coping with the stains and demands of the job. ID 569

A worsening situation

This survey describes a picture of increasing workload, falling morale and an accelerating workforce crisis. Since the initial survey in 2014,16 GPs’ stated intention to retire in the next 2 years has increased significantly with 48.5% of respondents to the current survey stating that they planned to leave working in GP sooner than they had expected 2 years ago. A majority reported an increase in hours of work since the previous survey, reflecting increasing workload, despite only 2.7% having expressed a wish to increase hours in the previous survey. Almost all (97.5%) were experiencing increasing appointment numbers to meet patient demand and 69.0% to manage patient complexity. The number of GPs planning a reduction in clinical hours has also increased. Many GPs are working over 40 hours a week and some up to 70, and most reported a reduction in morale and job satisfaction. In general, GP principals reported working substantially longer hours than salaried GPs or locums. These findings are in line with national findings of increasing consultation rates, length and clinical workload.18

The reasons given for intending to leave are similar to those described in earlier surveys.7 8 10–13 19 Workload remains the dominant driver to leave. Respondents who described having recently reduced their hours of work were more likely to express an intention to leave than others, suggesting intentions are affected by the nature and intensity of the work, together with other factors such as morale and job satisfaction, rather than by the number of hours alone. Given that one of the main reasons why doctors choose careers in GP is in order to have a better work–life balance,20 this increasing workload may result in disillusionment, low morale and be contributing to the increasing number of GPs choosing to work as non-principals and working fewer sessions from early in their careers.9

The survey was commissioned in part to discover whether the findings in Dale et al 21 about the negative impact of appraisal and revalidation on the retention of GPs in the West Midlands was replicated in Wessex. The Appraisal Service is unique in NHS England in being directly commissioned from an educationally based provider and has a conscious ethos of trying to facilitate appraisals with a strong emphasis on the support of the individual doctor. Although revalidation was reported as a minor factor in the intention to leave clinical work, appraisal itself did not emerge as a demonstrable factor.

The study identified that GPs vary in their enthusiasm for, awareness of, and experience of, national initiatives that are aimed at addressing workforce issues. Investment in practice nursing, closer working with specialists (eg, phone and email advice lines), investment in technology and expansion of the GP workforce were the initiatives that were viewed as being likely to have greatest positive impact. However, there was a widespread view that there were too many initiatives and that these were often complex to access; they would prefer for the investment to go directly to practices to decide how best to support their working practices. Despite this, the response to individual initiatives is mostly positive, with the exception of PAs, video and e-consultations and STPs. GPs who had experience of an initiative tended to view it more positively than others, suggesting that familiarity may lead to GPs becoming more aware of, and perhaps less sceptical, of their potential benefits.

The negative response to PAs in somewhat surprising in the context of the positive responses to increased numbers of nurses, pharmacists and paramedics working in primary care. PA training programmes are becoming increasingly in number across the NHS, and hence there may be a need to manage expectations for this workforce, as previously described22 despite evidence to suggest they are well received by patients.23 24 The Roland report14 viewed multidisciplinarity as one of the key solutions to sustaining primary care, though concerns have been raised about loss of continuity of care16 and resultant reduction in patient satisfaction.25 Future GP roles within increasingly diverse teams may need redefining and there has been interesting in alternative models of care,26 such as the Nuka system in Alaska.27

The strongest negative response was to STPs. Considering these are the main vehicle by which the 5-year forward plan for GP is being driven and support closer working between health and social care,28 that so many GPs believe they may make things worse is of concern. Further research in this area would be beneficial to understanding why many GPs lack confidence in this area, and what may be needed to promote greater positivity.

While investment in technology was positively received, e-consulting and video consulting were perceived negatively. Pilot studies have suggested that e-consulting may increase workload and costs as well as reducing patient satisfaction.29–31

Expansion of the general practice workforce remains a high priority to GPs. This has been recognised as an issue at governmental level, however, the response of increasing medical student numbers will not start to impact until 2028 at the earliest.32 An International GP recruitment programme has been set up,33 initially targeting GPs from the European Economic Area; however, there are concerns that uncertainties surrounding Brexit have impact on its success, and may result in EEA GPs currently working in the UK returning home.34

Perhaps the most interesting aspect of the survey was GPs’ views on what would improve general practice. More funding was the strongest theme, particularly for increasing the size of the workforce, both of GPs and other health professionals. This would enable a more manageable and sustainable workload, including longer appointments, so helping to reduce the risk of burn-out.35 Increasing financial demands including rising indemnity payments were also of concern, and there was enthusiasm for a national indemnity scheme. Reducing bureaucracy was identified, but less strongly than in previous surveys,9 possibly reflecting the reduction in incentive-related workstreams, the clinical value of which is now questioned.36 It is possible that the negative response to STPs relates to increased perceived bureaucracy. A number of innovative, more strategic suggestions emerged which may be worthy of further consideration.

Strengths and limitations

This study provides further evidence of the unfolding GP workforce crisis in England. A particular strength is that it demonstrates how attitudes are changing over recent years. However, the extent to which the findings could be compared with the earlier survey was limited due to privacy restrictions. There were some differences in characteristics between the two surveys (eg, respondents to the current survey were slightly older and were less likely to be GP principals) which need to be considered when interpreting comparisons. However, the difference in age profile was insufficient to account for the shift towards seeking earlier retirement.

The survey’s focus on how the crisis might be addressed is a strength, with the study providing evidence of the impact that national initiatives are felt to be having. The response rate was good for this type of survey; the questionnaire was quite lengthy and there was no incentive to support participation. The extent to which participants wrote free-text comments reflects the importance placed on this topic by GPs and added significant depth to the findings. However, it is likely that those who feel most strongly about their workloads either might have selectively responded to the questionnaire, or alternatively felt too busy and stressed to add completing a survey to their workload. Though this is in inevitable with this sort of study, it is a limitation in terms of drawing conclusions from the quantitative findings. While the findings are limited to a single region in England, they are reflective of views that have been expressed in other recent GP surveys and so are likely to have applicability across the NHS.