Investigating the contribution of physician assistants to primary care in England: A mixed methods studs

Developments in the UK general practice workforce

There are about 10,000 general practices in the UK and, as a sector, these have been characterised by entrepreneurial developments in service delivery as well as in practice organisational forms. 17,18 Many of these developments have arisen through individual enthusiasms and values19,20 and others have developed in response to incentives and health professional labour market fluctuations. 21,22 The flexibilities created both by the new General Medical Services Contract23 and other forms of contracts for primary care24,25 created opportunities for different organisational forms and skill mixes appropriate to the needs of patient populations and to the provision of out-of-hours and extended hours services. GPs, as both health-service professionals and employers, exemplify the human resource policy stream articulated first in the human resource strategy for the NHS Improvement Plan in 2002: ‘more staff working differently‘. 26 The GP workforce has, in the last 10 years, both grown and diversified, with increased numbers of part-time and salaried, rather than partner, GPs. 27–30 However, it should be noted that shortages of GPs and their distribution are not uniform across the country. For example, the number of GPs per 10,000 population in England in 2011 ranged from 6.3 in the east of England region to 7.7 in the south-west region. 27 There have also been changes in the types and volume of consultations undertaken in general practice, with shifts of responsibilities from secondary care to primary care, for example in chronic diseases management and the development of GPs with special clinical interests. 31

These shifts have been accompanied by GPs moving to employ a wider skill mix within general practice teams. 29 Such a development is evident in the growth over the past 25 years in numbers of administrative staff, nurses and others providing direct care. 27,28 Current statistics for general practice staff in England are illustrated in Figure 1.

FIGURE 1.

Composition of general practice workforce in England, 2011. FTE, full-time equivalent. Data source: NHS Information Centre. 32

In the UK, the solo GP with only a receptionist is an organisational model of the past. 33 However, debates continue within the profession of general practice regarding the appropriate mix of staff and skills for a general practice team. 20,34 It is evident from annual surveys in general practice in Scotland that nurses, together with direct care staff such as health-care assistants (HCAs), account for increasing volumes of patient consultations (30% in 2011). 35 Analysis indicates that these are not consultations in which patients present with undifferentiated diagnosis but are follow-up consultations for monitoring or review activities. 35 This suggests that the nursing and direct care staff are both supplementing the work of the doctor and undertaking delegated tasks. Outside general practice itself there are suggestions, such as from a HM Treasury-commissioned report, that NPs could provide a form of substitution, undertaking at least 25% of the work undertaken by GPs. 36 A Cochrane Review published in the same year37 suggested that the evidence was more nuanced. Previous systematic reviews indicated that patient health outcomes were broadly the same for NP and doctor consultations but that patient satisfaction was higher, the length of consultation was longer and the rate of investigative tests was higher for nurse consultations than for consultations with doctors. 38,39 The Cochrane Review excluded observational studies and also separated analysis between those studies in which the nurses were taking ongoing responsibility for chronic disease management and those in which they substituted for the doctor in first-contact, urgent consultations. 37 It was reported that in the studies of ongoing chronic disease management ‘there were no appreciable differences between doctors and nurses in health outcomes for patients, process of care, resource utilisation or cost’. 37 In the five included studies of first-contact, urgent consultations, broadly similar health outcomes for patients were reported for nurses and doctors. Differences reported between the two were lengthier nurse consultations, greater rates of recall by nurses, but higher satisfaction scores for nurse consultations than for doctor consultations. 37 The authors reported that the ‘impact on physician workload and direct cost of care was variable’. 37

By the turn of the twenty-first century there was a policy impetus for a redesigned workforce, particularly in primary care and first-contact services,21 an acute shortage of nurses and doctors applying to work in primary care and encouragement to develop and test new roles from a national body for service innovation known as the NHS Modernisation Agency. 40 At this point, some GPs with others who were looking for workforce solutions began to consider the contribution that another type of ‘mid-level’ practitioner, the PA, working in the USA, could make to the NHS. Before examining this occupational group in the UK, the following section provides contextual information on PAs in the USA and other countries.

Physician assistants

Physician assistants were introduced as a new professional group in the USA in the 1960s in response to medical shortages and misdistribution. 41 PAs are ‘graduates of an accredited PA educational program who are nationally certified and state licensed to practice medicine under physician supervision’. 42 The 170 educational programmes in the USA last an average of 27 months and require applicants to have qualifications in basic science and behavioural sciences and experience of health care. 42 The courses are based on medical education programmes and include 2000 hours of clinical experience practice. 42 Each qualified PA is required to undertake continuing medical education and recertification every 6 years. 42 PAs undertake physical examinations, investigations, diagnosis, treatment, and prescribing within their scope of practice as agreed with their supervising doctor. 43 They have to be registered in the state in which they want to work, and each state has separate regulations and limitations on the prescribing authority of PAs. There are similar numbers of PAs44 and NPs45 in the USA, and about one-quarter of PAs work in family medicine settings46 (Table 1).

Studies have demonstrated that PAs, as an occupational group in the USA, provide equivalent and safe care to physicians for the case mix they attend and are acceptable to patients. 47,48 PAs have been shown to have the same task productivity of between 50% and 80% of a physician, dependent on practice setting and experience of the PA, with attendant cost savings to the organisation dependent on the remuneration levels of both the PAs and the physicians they are replacing. 49–51

A number of other countries have PAs in their health-care systems and have developed PA educational programmes. Hooker et al. 52 provide an overview of international developments. Canada introduced PAs in the military in 1984 and in the civilian health service in 1999. Taiwan (Province of China) converted a group of nurses to PAs in 1997 and introduced a PA programme, as did the Netherlands in 2001. The health services in Australia ran pilot projects in rural and remote areas in Queensland starting in 2009, developed the first PA education programme in 201053 and are looking to develop this workforce further. 54 New Zealand has also run a demonstration project and is about to commence a second in rural areas and in primary care. 55 The International Association of PA Educators reported that in 2012 there were 12 countries with PA education programmes: Australia, Canada, England, Germany, Ghana, India, Kenya, the Netherlands, Saudi Arabia, Scotland, South Africa and the USA. 56 Key issues in all countries have been how to achieve recognised accreditation and regulation mechanisms as well as authority for prescribing. 52 The earliest record of GPs in the UK exploring the use of PAs was in 1980. 57 We will now describe the UK experience of PAs from 2000 to 2009.

Physician assistants in the UK

In undertaking a review of the literature, it became evident that in the UK setting the term ‘physician assistant’ was being used in two ways. The first was by some hospital trusts in the UK to describe medical technicians with national vocational qualifications (NVQs) at level 2 or 3. The second was in relation to PAs either trained in the USA or trained in the UK at postgraduate level to the American model. This study is concerned with the second group.

Under the auspices of the policy for the modernisation of the workforce, many developments in new roles were encouraged in the first decade of the century. Small-scale, local pilots of training for roles such as PAs (although often under different role titles) were developed in many parts of the country. 58 The central government health department interest in exploring the utilisation of PAs was evident in two ways. The first was the support of the development of the anaesthetist practitioner programme, later PAs (Anaesthesia), with the Royal College of Anaesthetists in 2003. 59 The second was the funding of pilot projects deploying PAs in support of a wider range of medical teams. Two large-scale pilot projects were funded from central health departments in which USA-trained PAs were employed in a variety of health-care settings: primary care, emergency departments, intermediate care, orthopaedics and surgical specialties in England (2002–5)60 and Scotland (2005–8). 61 The pilot in England arose specifically out of the difficulties of recruiting GPs to deprived communities and the interest in the PA role as a potential solution from a nurse partner in a personal medical services practice, some local GPs and the local primary care trust managers. 60 The evaluations reported that they were very well received by patients, received without problems by other professionals, worked at the level of either a NP or a doctor in training and were reported to be safe in clinical practice. Caveats were noted; for example, demands on mentors were higher than anticipated and a tendency towards slightly longer consultations was reported. 60,61 The evaluation study in England reported on eight USA-trained, experienced PAs working in GP practices. 60 Mixed methods were used including interviews, electronic appointment data and PA- and supervisor-completed forms of types of patients seen and their clinical activities over specified weeks. The PAs were found to provide a similar number of consultations a day (an average of 16.5 per day) to GPs (17 per day) and see, broadly, the same types of patients, with variations between individuals according to prior experience, subspecialisation and practice organisation. 62 Issues that were reported included a period of acclimatisation to the NHS and medical practice in the UK being necessary for American PAs, and a lack of prescribing rights resulting in a greater demand on supervising GP time than would be usual in the USA. 62 It was also noted that the PAs expanded the scope of types of services offered in some practices and allowed others to increase their numbers of registered patients. 62 The Scottish pilot evaluated supernumerary PA posts, of which five were in general practice. Mixed methods were used, including interviews with staff and patients and self-completion work activity by the PAs and events diaries by the PAs and their supervisors. 61 In the primary care setting, the findings concurred with the reports of the English pilot. 63 In their conclusions, both evaluations pointed to the need for research that considered cost-effectiveness more systematically. 60,61

Concurrent with these pilots, PA educational programmes were initiated62 and then further developed by the Department of Health, involving the Royal Colleges of Physicians and General Practitioners. Together these organisations have agreed a competency and curriculum framework for PAs’ education,64 modelled closely on that of the USA. By 2009, postgraduate diploma courses for PAs had commenced in four English universities and a further three trained PAs in the specialty of anaesthesia only. 59 The first of their students had graduated by 2009. These developments were supported by national PA networks and structures. The PAs in anaesthesia established an Association of Physicians’ Assistants (Anaesthesia), which has an agreed curriculum and competency framework with the Royal College of Anaesthetists and the Association of Anaesthetists of Great Britain and Ireland, and holds a voluntary managed register of qualified PAs (Anaesthesia). 65 The UK Association of Physician Assistants was established in 2005. 66 Together with the UK and Ireland Universities Board for Physician Assistant Education (UKIUBPAE), it established the Physician Assistant Managed Voluntary Register (PA MVR) in 2010. 67 The Department of Health agreed an updated version of the competency and curriculum framework64 to reflect this last development. 68

Although both the medical and the nursing professions have in the past been hostile to the introduction of the PA role,59,69,70 there has been increasing acceptance that the addition of PAs to multidisciplinary teams, such as first-contact services, can enhance the delivery of care. 71 At the same time, the number of general practices employing PAs, although small, is growing. 72 In an interview study of 20 GPs, respondents reported that they were motivated to employ PAs in order to increase the practice capacity to manage patient demand within government targets for access, while considering value for money and broadening the skill mix in their teams. 72 The GPs reported that the PAs were primarily being used to see ‘same-day’ and ‘urgent’ appointment patients, although some had also forward-booked appointment clinics and chronic disease management clinics, and some undertook home visits. That study did not quantify or collect details on the work activity of PAs. 72

It was evident that there was growing interest from potential employers in PAs and a potential supply of UK-trained PAs. The UK evidence was limited in 2009 as to the effect of introducing PA personnel into general practice on issues of patient outcomes, service efficiency and safety. This study aimed to address the knowledge gap and provide information to inform decision-makers in health services, commissioning and higher education institution (HEI) settings.

The review of evidence

In this element we undertook three review activities. The first was a review of empirical evidence using a rapid review methodology76 for empirical evidence of the deployment and impact of PAs (research questions 1 and 2; see Chapter 1, Study aims and objectives). The second was a narrative review77 of published commentaries and opinions of the PA role in the UK setting. The third was a narrative review77 of national and international policy and developments on health-care workforce planning in relation to PAs and primary care. The last two activities addressed the question as to factors supporting or inhibiting the presence of PAs in the primary care workforce.

Scoping survey

A scoping survey74 was undertaken to address research question 3: what supports or inhibits the use and development of PAs in the primary care? (See Chapter 1, Study aims and objectives. ) The method is one element of a scoping review74 and allows for data to be gathered from those knowledgeable in the issue of interest, providing a range of perspectives to ensure breadth in the analysis. 84 It is an approach that has been widely used, most recently in exploring European perspectives on human health resource issues. 85 A purposive sample of key informants at national and regional levels in England and Wales was approached to participate. These were individuals at a senior level from central departments of health, professional organisations, regulatory organisations, patient organisations, and commissioning bodies who are concerned with the development of primary care services, workforce development and health professional regulation. At the point of the design of the study (2009) the Department of Health and Social Services (Wales), unlike the Departments of Health for England and Scotland, had not explored the use of PAs. In order to consider diversity in national environments, individuals were approached to participate in England and Wales. Individuals were identified for approach through an internet search of public domain organisational sites, suggestions of the study advisory group and members of the research team, and snowball techniques from participants. An aide-memoire was developed of topic areas, based on the issues of interest, to be used in semi-structured interviews83 which were offered either face to face or by telephone, dependent on participant preference. The aide-memoire explored knowledge of mid-level practitioners, PAs and perceptions of the factors that currently and will in the future support or inhibit their use and development in the UK settings. Notes were taken during the interview and, with permission, digitally recorded. These were transcribed, anonymised and the recordings deleted. The transcriptions and notes were analysed using the constant comparison method86 by two researchers (WC and VMD) independently, compared and further analysis confirmed with a third researcher (JG). The analysis of the early interviews assisted in identifying further stakeholder groups from whom views were sought. This element of the study was deemed service evaluation by the NHS ethics query service and as such not for research ethics review. It conformed to all university requirements.

Survey of current use of physician assistants in general practices in England

An online, anonymous, self-completion survey was developed and piloted with two PAs who had previously worked in general practice in England. It posed questions of career history, work environment, work activities and estimates of volumes of work. PAs known to working in primary care were invited to participate in 2010 through e-mails sent out by the UK Association of Physician Assistants (UKAPA), course leaders of PA courses in English universities and practices identified as employing PAs in previous research. 72 All invitation e-mails included a link to the survey website and requested forward distribution to any other PAs thought to be working in primary care. UKAPA estimated that there were about 25 of their member PAs employed in primary care. The electronic survey was open for responses for 7 weeks (7 October to 30 November 2010). It addressed three areas: the activities and role of PAs in the primary care team, their primary care setting and their qualifications. Those interested in volunteering for the second phase of the study were given an opportunity to give their contact details at the end of the survey. The survey did not require NHS ethical review and fulfilled all university requirements for the ethical conduct of research. Descriptive analysis of the data was undertaken using Statistical Product and Service Solutions (SPSS) (IBM Corporation, Armonk, NY, USA).

Semi-structured interviews with practice staff

Semi-structured interviews83 were undertaken with PAs, GPs, practice managers, receptionists and practice nursing staff. A purposive sample in each practice was approached, provided with information and invited to participate. Aide-memoires for each professional group were developed addressing the questions (a), (b), (d) and (e), above, but allowing flexibility to explore new issues or ideas raised in the interview. GPs, PAs and nurses were also asked to describe their responsibilities and activities within the practice. In addition, practice managers and GPs were invited to provide any practice documents that would provide the research team with information about the practice, its working practices, services provided, staffing and planning. If documents on staffing were not available, the practice managers were asked to provide this information verbally. The interviews were conducted face to face or by telephone as preferred by the participant, digitally recorded with consent, transcribed and erased after analysis. Thematic analysis was undertaken88 initially by one researcher (RL), with further analysis by two other researchers (VMD and MH) and interpretative discussion with the wider research team. Documentary evidence on staffing was read and data extracted as relevant to providing contextual information for each practice.

Consultation record review and linked patient survey

The consultation record review and linked patient survey, which addressed question (b) above, was of patients presenting in urgent or same-day surgeries. This was the type of surgery that PAs were predominately working in as identified from the national survey (p. 9). Each practice identified time periods in summer and winter periods (known as reference periods) during 2011 and 2012 in which they were able to assist the research team in data gathering. The length of time periods were based on the practice’s estimation as to the number of surgeries required to reach the sample size. Each practice made staff and patients aware of the research during the reference periods in keeping with their usual practice, for example via notices at the reception desk.

Patient interviews

Patients who had consulted PAs and received a survey were also invited to participate in telephone or face-to-face interviews. This was designed to address question (c): how do patients understand the role of PAs and what is the experience of patients when consulting PAs? An aide-memoire was developed to explore issues not captured by the patient survey such as patient choice, understanding of the role of the PA, perceptions of the consultation with regard to communication and perceived expertise. A researcher (WC) contacted those who responded to the invitation as soon after the index consultation as possible. With consent, the interviews were digitally recorded, transcribed and analysed thematically88 by three researchers (LJ, MH, WC) using the NVivo software programme (QSR International Pty Ltd, Doncaster, Australia) (LJ, MH).

Observation of physician assistants and general practitioners

Patient consultations were video recorded, with consent of the practitioner and the patients, during one surgery per PA and GP. This element addressed questions (a), (b) and (e). We recorded GPs and PAs using two video channels, and assessed competence at consulting using the validated Leicester Assessment Package (LAP) rating scale. 105 Video recording is an established method of assessing the consultation106 and the use of two channels was a compromise between the multichannel recording we planned106 and what we could gain the NHS REC permission for.

Patients in the designated surgeries were provided with information and consent was sought in writing by a researcher in the waiting room before recording commenced. Recording was stopped at the end of each consultation. Patients were then asked whether they continued to give consent or would prefer that it was withdrawn and the recording deleted. Practitioners also reconsented to the inclusion of each recording or could ask for a recording to be withdrawn at the end of the surgery. The recordings were made using ALFA: a computer-mediated consultation observation technique developed to provide an analysable overview of the consultation. 107 Each consultation was assessed by two GPs independently using the LAP105 for assessing competence in general practice on dimensions of interviewing and history taking, patient management, problem solving and behaviour and relationship with patients.

The LAP for video consultation is a tool for measuring competence by direct observation. It was initially developed in the 1990s and demonstrated to have validity108–110 and reliability across levels of clinical competence. 107 Testing of reliability found a range of mean scores between 51.3% and 70.2% in one study, depending on levels of experience,110 and between 45% and 75% in another looking at the effect of interruption. 109 Although it has been criticised, no better package has been developed. 111,112 The LAP has also been used in a range of clinical contexts – assessment of cancer patients,112 the needs of caregivers113 and in the assessment of psychosexual problems. 114 This package has also been used in an international setting, where a 50% pass mark was included,115 for assessing practice nurses116 and medical students. 117

Grades were given by the four raters, using the seven main categories of the assessment package, and the results were then analysed. The seventh element of the package, the record, could not be rated as it could not be properly observed from the two-camera set-up used. Record keeping referred to how the health-care professional (HCP) interacted with the computer. It was not possible to view the detail of the individual record being collected by the HCP owing to the fact that only two channels of video were being recorded, and an additional channel is required for recording video directly from the HCP’s computer screen. Three of the seven categories – interview/history taking, patient management and problem solving – are given a double weighting, and count for 20% of overall marks (Table 4). When combined, this means that the LAP scale total is 100, allowing final results to be expressed as a mean score (out of 5 for each category) or be expressed as a total percentage score.

A workshop was held to discuss the findings with the raters (unblinded) which was video recorded and transcribed.

Observation was also undertaken in a sample of practice clinical meetings in which PAs and GPs were both present to understand more broadly how the PAs interact with the GPs in clinical decision-making in the practice. Permission and invitations was sought from the practice manager, the GPs and the PA. Permission and consent to observe was sought at each meeting that the research team was invited to. Field notes were made of only the activity and interaction of the team members. Analysis of these notes was used to complement data from the staff interviews (see Semistructured interviews with practice staff) and consultation video observations.

Work activities of physician assistants and general practitioners

Work activity diaries, using the 2006–7 general practice workload survey,118 were offered to PAs to complete for 1 week in the summer and winter. This addressed question (a), above. Early discussions with GPs made it evident that these diaries were unlikely to be completed by practice professionals other than the PAs and the data were therefore sought through interviews (see Semi-structured interviews with practice staff) and administrative systems. Data were entered onto a SPSS spreadsheet and analysed descriptively.

Economic analysis

The economic analysis was conducted at two levels:

1. The resource implications and costs of different team configurations were explored using practice-level data on staffing. Skill mixes, and the relative contribution of different practitioners, were compared within and between practices that do and do not employ PAs, after consideration of case mix differences. Nationally validated scales119 were used to calculate total human resource costs and costs per patient.

2. The impact of PAs, compared with GPs, on consultation processes (e.g. prescribing, referral and investigations), consultation outcomes (e.g. reattending, consultation satisfaction) and length, at individual patient level, as analysed in the clinical record element were incorporated in a high-level analysis of costs and effectiveness.

Ethics and research governance

Phase 2 of the study required a number of types of reviews and agreements. The research ethical review was undertaken by NHS Research Ethics Committee (REC) South East Coast – Surrey who gave approval but required that child patients were not included in some elements (REC reference number: 10/H1109/28). Local NHS primary care research governance approvals were obtained using different local processes from five different bodies. The procedures for local agreement of the UKCRN NHS service support costs to research in general practice varied in each area. Agreements were obtained for each participating general practice.

The case study settings

Twelve practices participated in the study in the south-west, east and south-east of England. They were diverse in the sociodemographic features of their locale, the number of practice partners and the size of their registered patient population. These data are presented within ranges, to ensure anonymity for the practices, in Table 5.

The report now turns to presenting the findings.

The systematic review of empirical evidence

Of the 2167 publications identified and screened, 49 met the inclusion criteria49,62,127–161 as presented in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA)162 diagram in Figure 2.

FIGURE 2.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA)162 flow chart of the review. Adapted from Halter et al. 176

Of these included studies, 46 were from the USA, and one each was from the UK,62 Australia,163 and the Netherlands. 164 The heterogeneity of the types of studies (Table 6) and their outcome measures mean that any meta-analysis was not possible and a narrative review is presented. The overall judgement of the quality of the research was that it was weak to moderate with few studies providing comparative data about other occupational groups.

Evidence regarding structural aspects of PAs in primary care was found in 24 studies62,127–149 ranging in date from the late 1970s to 2007. These studies focused predominantly on PAs as a human resource in primary care, investigating questions as to the numbers employed in primary care, the willingness of physicians and managers to employ PAs, perceptions of barriers to employment and issues in retention rates. In the USA, the numbers of PAs are seen to have grown over time. Although studies vary in their findings, several suggest that approximately half of those with PA qualifications work in primary care. 128–132,139 Support for the concept is mostly reported to be high among family doctors,62,142–148 with the potential for releasing the doctors’ time to allow them to focus on more complex patients (and thereby increase practice productivity)143–146 being one of the most prominent reasons, alongside perceptions of patient satisfaction. 146,147 Retention is not considered problematic. 140,141

Aspects of the process of care were reported in 24 studies49,62,134,137,140,144,147,151–161,164–168 ranging from the late 1970s to 2009. The majority of these studies describe the work activities of the PA and the patient groups who consult them. The greater part of the PA work is reported to be in consultations with patients with acute presentations. 140,150,151,153–155,158,160 Two studies suggest that doctors who are in the same primary care facility as PAs attend more patients with chronic conditions49,153 and one suggests that the doctors attend more patients with more serious problems. 140 The evidence indicates that, while PAs can be consulted by patients of all types, primary care facilities have systems which direct particular groups of patients to the PA. 156 The activity level of PAs’ work is reported in a number of studies but without conclusive evidence. Varying levels of PAs’ productivity in patient throughput have been reported compared with those of family physicians, suggesting that there is near equivalency62 and up to 60% physician whole-time equivalence. 164 PAs are consistently reported to require supervision from physicians, although this is low in actual patient care episodes154,165,166 apart from when PAs are not afforded prescribing rights. 62 Some evidence suggests that the employment of PAs provides greater capacity and efficiency in the provision of primary care. 62,149,153,164 Seven papers49,154,155,157,158,167,168 report on investigations in the USA about the issue of cost and the impact on practice finances. Only one provides data after the year 2000. 49 Again, no conclusive evidence is provided, with four suggesting that the employment is not financially advantageous for the practice49,155,156,168 and four suggesting otherwise. 38,154,158,167

Evidence about the outcomes of care was found in 14 papers. 140,141,144,154,155,160,163,168–173 Overall, PAs were reported as acceptable to patients,141,155,160,168–171 although this decreased in a rural setting in which patients were not offered the choice to see a doctor. 141 One study involving hypothetical scenarios suggested that patients were less willing to consult a PA in primary care if they had more complex conditions. 173 The evidence on technically appropriate care provided by PAs was weak and mainly from the 1970s. 140,144,154,155,168,169,174 While evidence was mainly positive regarding the appropriateness of care, some less favourable comparisons with other providers were also reported. 155,175 It is noteworthy that most of these studies were concerned with care for people with chronic conditions, for example diabetes; however, the evidence suggests that this patient group are in the minority in the PA workload.

Despite 40 years of studies, the evidence pertaining to PAs in family practice remains descriptive and weak, accentuated by poor reporting. While the review provides some evidence of the consequences of changing from physicians to PAs, it also makes it clear that research questions remain which would benefit from country-specific, good-quality investigation.

The findings of this systematic review are reported in full elsewhere. 176

Review of commentaries

This section reports on the findings of the review of commentaries and published opinions. This was one element of the study that contributed to investigating the question as to what supports or inhibits the development of PAs in primary care in the UK. The review was undertaken as reported earlier (see Chapter 2, The review of evidence). One hundred and seventeen published (print and electronic) commentaries and opinions were found that met the criteria for the review and these are listed in Appendix 9. They were found in sources with public, medical and nursing audiences (Table 7).

The commentaries ranged in date from 1980 to 2012, with the greatest number published between 2004 and 2006 (Table 8). This was the period in which a Department of Health-funded pilot project was being conducted in England. 60

Journalists were the most frequent type of author (n = 28), followed by medical staff (n = 24), including those with clinical backgrounds and senior medical practitioners, and by academics (n = 20). A notable absence was a patient or public perspective. Only one member of the public177,178 pointed out that the public were not asked for their views and questioned the need for new roles in the NHS. Many of the commentaries were unclear as to the competencies of a PA. In some, the author confused a PA trained in the USA model with a similarly named role in the UK in which individuals are trained to undertake only technical tasks such as taking bloods under the direction of a doctor.

Fifty-one of the commentaries gave a positive view of PAs, 24 gave negative opinions and the remainder were neutral, often suggesting that more information was required to help inform opinion. The positive opinions were authored for the most part by senior medical practitioners and academics. The doctors were mainly from the acute setting and, either from direct experience or from consideration of their current staffing issues, could envisage a place for PAs in their teams. Some of the more recent opinions relate to the potential for PAs to provide solutions to contemporary health policy requirements such as those found in the Quality, Innovation, Productivity and Prevention programme174,179 and as support to GPs in the new commissioning structure. 180 The majority of the negative opinions were authored by GPs and nurses. None of these had direct experience of PAs. Issues raised included the transferability of a US model to a UK setting; confusion for the public when faced with multiple roles; concern that PAs were not cost-effective in general practice when compared with GPs; and a viewpoint that nurses and NPs fulfilled this role in the UK health-care workforce and offered greater value to patients than another professional trained in the medical model. Conflicting views were authored from junior doctors; some reported that PAs made their working lives better as they could delegate routine tasks, while others questioned whether or not they would be in competition for jobs for which PAs would be seen as preferable by managers because they were cheaper.

There was little citation of empirical publications to support opinions in the commentaries, with most referring to other opinion pieces. It is possible to detect a trend over time to more positive, or at least neutral, opinions as more direct experiences of PAs in the UK were reported.

We turn now to the findings from reviewing the workforce policy documents and associated material.

Review of workforce policies and developments

One hundred and thirty-three documents were identified. Those concerned with international developments regarding PAs in primary care informed the context for the study (reported in Chapter 1). Twenty documents were published after 2010 (the year of the election of the Coalition government) by central government departments or regional bodies responsible for NHS workforce and education planning. There was a lack of reference to PAs in all of the English policy documents and, similarly, in the House of Commons Health Committee Report into the workforce and education. 181 At a local level we did identify a few NHS trusts in England with plans to employ PAs (e.g. see University Hospitals of Leicester NHS Trust Annual Operational Plan 2011–12182). The only mention of PAs in Welsh documents was in reference to the health workforce in rural areas, but no commitment to develop or support the role of PAs in this context was made. 183 In contrast, there was more evidence of recognition for PAs in Scotland where they were cited in central government policy documents both as an example of workforce efficiency184 and as a potential addition to the remote and rural health-care workforce. 185 To this end, NHS Grampian, as the lead NHS organisation for workforce development, was working actively with the University of Aberdeen to establish and recruit to a PA education programme. 185 There was little reference to PAs in other types of workforce review documents which were intended to inform decision-making bodies.

Support for PAs was, however, evident in the medical colleges. The support of the Royal College of Anaesthetists for PAs (A) was noted earlier in Chapter 1, as was the support of the Royal College of General Practitioners and the Royal College of Physicians (RCP) in establishing competency and curriculum frameworks. More recently, the Royal College of General Practitioners referred to PAs in their review of general practice34 and also included PAs in their membership via their General Practice Foundation. The Royal College of Physicians of Edinburgh also offered foundation membership to PAs. 186 PAs were also cited by the Director of the Medical Workforce Unit at the RCP187 as one of the solutions to the problems facing medical workforce planning, medical training and complying with the European Working Time Directive. 188

Against this backdrop of limited recognition in policy, we go on to report the findings from the scoping survey.

Health-care workforce changes

Health-care workforce redesigns and changes were seen as ongoing and inevitable by all respondents. They identified the current and future drivers which included the demographic and epidemiological changes in the population; the need to contain health-care costs and the impact of the general economic environment; the technological changes in medical treatments and communications; public and patient expectations in health-care delivery; and employment legislation and changes to professional education and regulatory requirements, specifically with regard to medicine and also the implementation of the European Working Time Directive. 188

Most informants had either observed or been involved in some way in shifts in work roles between occupational groups and envisaged this movement continuing. For example, the participants from patient organisations described the growth in employment of HCAs in NHS hospitals, who were undertaking what had been previously registered nurses’ activities. Some described the shift in terms of ‘the blurring of boundaries’ between occupational roles, while others described planning for some groups to gain competencies (assessed in some way) to take over roles or tasks of other occupational groups, as in this exemplar:

So we take the approach of developing registered professionals into, you know, roles that will support medics or even, you know, replace some of the work of the junior doctors because with less junior doctors being around, you know, we’ve got a real capacity issue so registered professionals are currently filling that gap.

Participant 17: regional workforce development organisation

Workforce changes in primary care

In considering additional factors that impacted on workforce developments in primary care, participants noted the policy drive to deliver more care in primary care settings, reduce unplanned hospital admissions and increase the availability and accessibility of primary care. In addition, one GP predicted changes in the actual GP workforce, for example increased part-time working, and their possible impact:

I think in general practice there’s going to be a much greater feminisation of the workforce, so I mean that’s going to cause a tension because of people, you know, working out of normal working hours and kind of family and caring commitments making that very difficult, so you might almost get two types of workforces, those who work evenings and nights and those who work, you know, main work days.

Participant 5: GP and primary care organisation

There seemed to be general agreement as to the direction of travel of developments in the workforce in primary care. These included (1) greater numbers of staff (medical, nursing, allied health professions, support staff and administrative staff) to provide the range of services moving from hospital to primary care settings and accessible for longer hours, 7 days a week, (2) increased availability in primary care of medical specialists in ambulatory sensitive conditions, (3) increased subspecialism by GPs, and (4) increased stratification in some occupational groups, for example more clinical nurse specialists and advanced NPs, fewer registered nurses and more HCAs.

Experience and knowledge of physician assistants

Aside from those involved in organisations concerned with PAs, about half of the participants had no direct experience of an aspect of the PAs movement in the UK. Those without prior knowledge, such as participants from patient organisations, could easily translate the concept and sought additional information to check their understanding. Of those who had direct experience, this ranged from involvement in early pilot projects of PAs or staff like them to consultations of professional bodies on the Department of Health-agreed national competencies for the PA programme,64 to engagement with provider organisations on the commissioning of places on PA programmes. This did not necessarily translate to knowledge of any current developments in the PA movement in the UK.

And you have to look back, not so long ago, but we had this thing called the ‘NHS Plan Modernisation Agency’ and with it, there were all these calls for workforce development. And one of the calls was for England to develop the role of physician assistant to increase capacity and capability within both secondary and primary care. Now, as a response to that, a working party was set up looking at the education programme that SHAs would commission from HEIs for the physician assistant, and I was on that working party . . . but I had assumed that we had very few out there now.

Participant 9: nursing profession organisation

Participants also commented that they thought there was little awareness or knowledge of PAs among the public or, indeed, professionals.

So you know if you talk about physician assistants I actually don’t think many people in this country really understand what they are.

Participant 15: GP and medical profession organisation

Factors influencing demand for and employment of physician assistants

Around half of the participants saw potential for the PA role in primary care, although there were caveats expressed, as outlined below. Other participants were more neutral in their opinions, requiring more information and evidence. None expressed a completely negative view of the potential for PAs to be employed in primary care settings.

For those who could envisage a role for PAs in primary care, that role tended to be one of triage or dealing with patients with minor, self-limiting illnesses in general practices and other first-contact settings such as walk-in centres. The explanations for this were that this was the area of patient pressure on the practices, but also that the role could free up the doctors’ time to deal with patients’ more complex problems or with other types of activities such as commissioning. Some participants reported that PAs were increasingly being employed in secondary care settings.

Something like seven out of ten of the PAs working in the UK now are in secondary care . . . something like 19 trusts are now employing PA’s across the UK. There are none in Wales that I know of and none in Northern Ireland but the vast majority are in England.

Participant 16: PA organisation

This echoed views reported to one of the research team members when interviewing senior staff involved in the New Zealand pilots of PAs (see Chapter 1).

The factors that participants considered were particularly likely to create demand for PAs were lack of medical applicants to work in particular areas or types of primary care services; the need to have continuity in medical team personnel; and the need to free medical personnel from the more routine aspects of patient care to either undertake their medical training or undertake other types of work. Some participants debated whether it was nurses at advanced level or clinical nurse specialists that were required or whether or not it was PAs. Most considered that they needed further information before taking a view, although one participant from a professional nursing organisation considered that the nurse added value in ways the PA would not.

Having had a long conversation with one of the PAs [at professional meeting] to actually do a direct comparison of what were the pros and cons of an ANP [advanced nurse practitioner], because they couldn’t understand what an ANP was and say ‘Well, why would you need ANPs, you’ve got physician’s assistants’ so we were like diametrically opposed and I’m going ‘Well, why do you want physician assistants because we can get that and more from an advanced nurse practitioner’.

Participant 11: nursing professional organisation

Participants from medical education and medical professional organisations could see particular problems for consultant teams created by the changes in postgraduate medical training, which roles such as PAs could address. One participant noted that in a LETB, representatives of trusts that had experience of training and employing PAs were voicing requirements for a greater supply of PAs.

. . . the consultants there who’ve hired now four physician assistants, they were losing training roles because their trainees couldn’t get in, because they were having to do the ward work and couldn’t get into surgery. With physician assistants in place they have the ward covered, they’re not hiring locums, they’re now getting more training posts because they can actually train those people.

Participant 24: medical education organisation

In contrast, in a different region of England, a participant from a workforce planning organisation reported a lack of interest by trusts in PAs specialising in anaesthesia.

Yeah, I mean we constantly look at the potential around PAs in the region and have had little uptake really locally. We’ve had a few events where we’ve explored what physician assistants are but none of the trusts have wanted to go forward with them.

Participant 17: regional workforce development organisation

Participants also raised questions and offered views as to why PAs might not be employed in primary care settings. Foremost were questions (not in any order) as to the cost and value for money compared with other staff groups, the lack of legislated authority for PAs to be an independent prescriber and the lack of oversight by a statutory regulatory body. Opinions seemed to vary as to whether or not there was a shortage of doctors applying to general practice to which PAs might offer an alternative but that may have reflected regional variations in settings. Participants had different views as to whether or not PAs represented value for money when compared with doctors. Most used the evidence from studies of NPs, reporting that compared with doctors (particularly in general practice) their consultations with patients were longer, resulting in no financial gain to the service. They considered that the evidence about PAs would be similar.

Practices have tried nurse practitioners and generally they take twice as long to see the patients and don’t come to the conclusion or the management as quickly. So you actually often end up when you look at the, and there’s been published papers to show it, when you look at the cost actually the nurse practitioner doesn’t become any cheaper.

Participant 15: GP and medical professional organisation

The participant from the patients’ organisation also raised issues of opportunity costs to patients in consulting a PA who could not prescribe.

It seems to me that if the physician assistant can’t actually prescribe then, you know, you’re duplicating . . . you’re repeating what has happened with, you know, some of the walk-in centres experience, where people have gone, you know, thinking I won't need to bother with the GP. They see a nurse at the end of which the nurse says, ‘Well I think you'd better go and see your own GP service.’ . . . So actually how much, how effective is delegation if people just bounce back through again because the GP’s the only one who can make a particular decision?

Participant 12: patient organisation

Issues of acceptance and patient preference were mentioned by a range of participants. A number cited experience and research that demonstrated patient preference for consulting with a doctor rather than a nurse, although some also pointed out that many patients, particularly those with long-term conditions, were also discerning as to the expertise level they required and found acceptable.

Concerns were also expressed that the PAs were not a professional group regulated by legislation, although, as one participant from a patients’ organisation pointed out, the existence of a regulatory body would not necessarily result in regular checks of individual competence. The absence of such a body was considered by some, but not all GPs, as a barrier to employment:

No, I didn’t realise that [there was no regulatory body for PAs], not in a million years would I employ someone.

Participant 5: GP and primary care organisation

Participants from the PA organisations reported the existence of a voluntary register for UK PAs which recorded individual PAs’ credentials. In addition, it was reported that UKAPA had tried to be included in the remit of the Health Professions Council but, following the government’s announcement not to regulate any further groups,191 this was not currently possible.

Finally, a number of participants made comment as to the risk-averse, essentially conservative attitudes of many NHS managers and employers towards innovation. In comparison, one GP participant noted how ‘nimble’ general practices were like small independent businesses in the speed at which making decisions and testing innovations could occur, particularly when compared with large NHS trusts. Others observed the need for evidence and information so that employers and commissioners could make decisions. Some participants considered more widely the factors in diffusion in innovation and offered a broader analysis that included a NHS cultural aversion to the importation of ideas from other countries such as the USA.

So, I think this is an opportunity to try and move some of these ideas. You know, they may not be right of course. So, ‘Do they feel right?’ but I don’t think physician assistants have ever had enough scope to run a NICE [National Institute for Health and Care Excellence] trial on their VFM [Value for Money] and impact. Which is probably part of the thing that drags you back down. It’s almost like you’d need to reach that tipping point even in terms of evidence or just acceptability to get it done. And the weird thing is that ‘and not invented here, culture’, [you] can’t point to 175,000 American physician assistants and say, ‘Oh yeah, it does work’ you know, cos it’s not here.

Participant 20: workforce development organisation and national and regional level

The scoping survey suggested that there was awareness of potential benefits of the role but also recognition of the current drawbacks and a desire for more information.

This section has reported on the findings of the scoping survey as to perceived factors influencing the use and development of PAs in primary care at the macro and meso levels. We now turn to the evidence as to the employment of PAs in England in primary care.

Physician assistant activities: an overview from the work diaries

The work diaries (confirmed in interviews with PAs) indicated that the majority of the PAs’ working hours were spent in providing clinical consultations (direct patient contact in face-to-face or telephone appointments), although the proportion of the time they reported on this ranged from 58% to 92% (Table 11). The majority of these hours were in same-day appointment sessions but PAs also undertook booked appointment sessions. These were mostly for the follow-up of patients with long-term conditions.

The non-clinical activities could be related to either patient consultations, such as arranging referrals or processing laboratory results, or other professional activities such as attending practice meetings or training days. The most frequently reported non-clinical activities by PAs were dealing with test results and referrals to other professionals. Some PAs also reported attending training and carrying out clinical governance activities and condition/case-specific duties (such as child protection work). It was evident that the wider the range of non-clinical activities, reported by individual PAs, the greater the proportion of time spent on these activities compared with clinical activities (as with PA 10) (see Table 11).

From the staff interviews, it was evident that each practice deployed the PAs according to a number of parameters. These included deployment that fitted with the practice organisational requirements, for example to telephone triage in a practice that used that system, or by clinical competence as judged by the senior or lead GP, as in this exemplar:

[The PAs] don’t usually see under-1s but depends on [their] experience.

GP 11–5

This is further explored in the following sections, where the variety in the scope of practice reported to be undertaken by the individual PAs is clear.

Same-day appointment clinical consultations

The deployment of PAs in the practices was primarily to provide clinician time for same-day/urgent appointments. The range of scope of practice was evident in the manner in which same-day appointment patients were assigned to a PA or to a GP. In some practices, it was evident that the receptionists treated PAs as the doctors assigned to that type of surgery, as noted here:

He [the PA] sees a surgery of patients morning and afternoon every day, which are almost entirely unselected. We have selected out under ones because he is not trained for those, but other than that he sees the full range of problems that present.

GP 1–2

In other practices, more nuanced systems had been developed so that the GPs and PA were seeing different types of patients, with the GP more likely to see either the more complex patients or those with more chronic conditions, as in this exemplar:

So the doctors and the physician assistant, the nurses between them have developed a, like a list, a triage type list. So if someone rings up and says ‘I’ve got this, that or the other’, they [the reception staff] can look on the list and decide whether it should see a doctor or a physician assistant or a nurse practitioners, depending on who’s on. So they’re a sort of triage but not in-depth triage, just pointed in the right direction.

Practice manager 6–4

The practices varied in the length of the appointment slots they assigned to PAs. Most – but not all – assigned longer slots (15 minutes) to PAs than to GPs (10 minutes), but often shorter appointment times than those allocated for nurses.

Chronic disease management consultations

The extent to which PAs were involved in the ongoing management of patients with chronic diseases also varied among practices. For example, in one practice the PA undertook the diabetic review clinic, among other chronic condition review clinics. Another PA estimated that about 50% of the patients he saw were ‘regular routine follow-up of chronic patients’ (PA 5–3).

Some of the practices had nuanced ways of separating the work the PA undertook in chronic disease management from that of the nurses in the practice. For example, in most practices the nurses were described as not seeing patients with depression or other mental health problems or with musculoskeletal problems, whereas the PAs would attend to those types of patients. In general, the division of labour appeared to be based on clinical expertise and confidence, as in this example:

When the physician assistant that we have here, when she started a lot of my chronic disease patients [were] then allocated to [the PA], especially COPD, she does spirometry, which she’s been trained for . . .

Practice nurse 10–3

It was observed by a number of participants that the difference between nurses and PAs in their chronic disease management was often recognisable through the point at which the patient was referred back to the doctor. PAs were described as often managing more problems and transition points, whereas the nurses referred back for any problem or change. The differences in activity between the PA and the NP were exemplified in this description:

She [the PA] does all the things the nurse practitioner will do on the day but she also does a couple of other things because she’s covered the training. So she is happy to see straightforward depressions, which the nurses, none of our nurse are prepared to at present . . . And she is happy to see back pains also which the nurses haven’t covered so they don’t do that. Another difference is the PA will do some of her own referrals [to secondary care] whereas the nurse practitioner would refer the patient to the doctor who would be responsible for doing the referral.

GP 6–2

More evident in the descriptions were situations in which PAs and nurses flexibly covered the work required in the practice as the demands of each day unfolded. These were reported by PAs, nurses and practice managers; for example, PAs might cover the work of nurses absent through sickness, while one nurse described assisting with booked chronic disease management review patients in the absence of a PA.

Income-generating work

In addition, some practices were also developing their PA to undertake work that contributed specifically to the practice income as part of either Quality and Outcomes Framework (QOF)93 indicators, such as maintaining the register of people with learning disabilities, or clinical activities that attracted specific payments under local or nationally agreed enhanced service elements of the General Medical Services (GMS) or Personal Medical Services (PMS) contract,23 for example initiation of insulin in primary care. Each of the activities referred to by the PA in this exemplar is an activity associated with specific payments:

What else do I do? I’m responsible for the warfarin patients, I do all the bridging for surgery [changing the warfarin pre-surgery to another drug and restarting post a surgical operation]. I do contraceptive implants and there’s only one or two in the practice who do that so that’s quite a big role . . . and I do a lot of paperwork, so I run the palliative care meetings 3-monthly . . . the GP would normally do that but I’ve taken that off their hands. And I’m now starting to do initiation of insulin in primary care. We had one GP that was doing that but he retired.

PA 6–3

Authority and lack of authority to order tests, refer and prescribe

There was a reported difference between PAs in different practices and between PAs and nurses as to the recognised authority to order pathology and radiological tests and also to make referrals to secondary care. In general, the PAs were described as having been given more authority to order tests and make referrals to secondary care (although some needed countersigning in some areas by the GP) than the nurses, who mainly described usual practice as referral back to the GP, as described in this exemplar:

Can you authorise an X-ray referral?

No that goes on our prescription board [a way of leaving items such as prescriptions for the doctor to sign] and the doctor will sign it. Obviously it is a trust thing, they know what I can do and trust me.

And if they needed investigations of any other kind, would there be something you authorise or . . .?

Bloods. I would obviously do that myself and authorise myself. I think X-rays, most other things, scans and things would be the doctor or PA.

And if they need a hospital referral?

Yeah doctor or PA because they wouldn’t take any notice of me at the hospital.

Physician assistants are not authorised to prescribe medicines in the UK. The lack of ability to prescribe was reported to cause frustration and additional time to the PAs, the GPs and the patients. Each practice had developed systems to try to minimise the delay caused by getting the GP to sign prescriptions. This varied from a system of knocking on the GPs’ consulting room door, messaging by computer, hovering in corridors to speak to the next available GP, and leaving the prescription in a designated place for the next available doctor to review the notes and sign, to offering to send the signed prescription to the local pharmacist later for collection. It was evident that the GPs developed trust in the clinical competency of the PAs over time, which had the effect of reducing the amount of time they spent reviewing consultations before signing prescriptions, as illustrated in this exemplar:

When I first qualified and also when any new doctor starts there’ll be a period where when they are signing my prescription you would tend to give them a lot more information about what you are doing . . . explain a bit more what it’s about, because they have got to learn to trust you and understand it, because ultimately they are signing the prescription, . . . so it’s just about building up a trust and an understanding of your competencies which is fair enough.

PA 10–2

In the PA-employing practices, the nurses with the qualifications to prescribe medications were in the minority. Consequently, most of the nurses, whether with a title of nurse practitioner, senior practice nurse or practice nurse, were using the same systems as the PA to obtain GP signatures on patient prescriptions.

Consultation records sample numbers

The sample initially identified for inclusion in the study consisted of 2260 patients, reduced to 2086 after exclusions, for analysis. Figure 3 presents a flow chart of the inclusion and exclusion of the patient consultations forming the study sample.

FIGURE 3.

Flow chart of inclusion and exclusion to the consultation records and patient survey.

In total, 932 consultations with PAs and 1154 with GPs were included for analysis. The number of patient consultations per practice ranged from 90 to 255 among the PA practices and from 136 to 304 among the GP practices.

These differences reflect the amount of time the practice could commit to the study, the proportion of patients seen who were under the age of 16 years and the proportion of eligible patients to whom a survey was offered.

Work activity of physician assistants and general practitioners in the consultation record sample

The work activity of PAs and GPs in the sample for the index events is described in this section through the demography of the consulting patients and a characterisation of the health problems of the sample.

Characterisation of the health problems of the consultation record sample

The sample’s health problems are characterised through the number of problems the patient presented with at the index event, the patient being registered on one or more QOF93 disease registers, the number of repeat prescriptions issued to the patient in the year preceding the index event, the number of attendances at the general practice in the 3 months preceding the index event, and a classification of the severity and acuity of the presenting patient as a whole through a combination of their presenting problems (building on de Jong et al. 96) and their presence on QOF registers. This is referred to as the ‘PA study condition classification’.

The number of problems the patient presented with and which were recorded in the index event ranged from one to eight in both PA and GP patient consultations, though 73.9% of the sample presented with just one problem and another 17.9% with two problems. The remainder of the patients, in steadily diminishing numbers (each with less than 5% of the sample), had from three to eight different problems recorded in the index consultation (Table 17). No difference is observed between the PA and GP consultations.

Patients in the sample presented with being on up to seven QOF registers, though 60.8% were on no registers, another 20.5% were registered on just one, and the remainder, in much smaller numbers (each with less than 10% of the sample), were on between two and seven QOF registers (see Table 17).

The majority of the patients in the sample had between zero (27.1%) and three (11.2%) visits to the practice in the previous 3 months (see Table 17). The remaining upper quartile of patients was spread, in steadily decreasing numbers, between 4 and 30 previous visits to the practice.

The number of repeat prescriptions patients were recorded as having been prescribed in the year preceding the index event ranged from none for 45.9% of the sample, one for 14.2% and two for 10.0% up to 26 and 27 repeat prescriptions for one person each in the sample (see Table 17).

Descriptively, GPs appear to be consulted by patients whose average number of QOF registers is higher, who have attended the practice more often in the 3 months preceding the index event and who have had a higher number of repeat prescriptions in the year preceding the index event than those patients consulting PAs.

The ‘PA study condition classification’ sought to address severity, acuity (medical immediacy) and complexity of the presenting complaint(s) in the context of other health characteristic information about the patient, and placed some patients in each classification of acute, chronic, minor/symptoms, prevention and processes for both PAs and GPs. One thousand, one hundred and forty-seven Read code descriptors existed for the presenting problems of the 2086 patients at their index events. Nineteen per cent (n = 213) of these were listed by de Jong et al. 96 in their classification of acute or minor problems. The largest grouping was the minor/symptoms category (Table 18). Significant differences were found between patients consulting PAs and GPs, with PAs tending to see more patients with problems classified as ‘minor/symptoms’ than GPs, while the GPs saw more patients classified as ‘chronic’ or ‘acute’ (χ² = 34.51; df 4; p < 0.001).

The ‘PA study condition classification’ appears to be correlated with patient age. The patients classified as having a ‘chronic’, ‘acute’ or ‘process’ presentation have a higher mean age than those classified as having a ‘minor/symptoms’ or ‘prevention’ presentation (Table 19).

These findings suggested that all analyses of process and outcome would benefit from statistical adjustment for the ‘PA study condition classification’ and for the patient’s age as a baseline.

Consultation process outcomes

The process outcomes within the consultation have been measured as follows: the number of procedures carried out in the consultation; number of investigations or tests carried out or ordered; number of prescriptions from the index event; number of referrals made to other services; general advice given; advice on the management of medicines; advice on the use of over the counter medicines; and the issue of Fitness to Work certificates.

Numbers of patient consultations recording process outcomes

The number of patient consultations with each these process measures recorded is shown in Table 20.

The data show that:

• Very few procedures were carried out at the index consultation by either PAs or GPs (mean 0.1, median 0 for PAs; mean 0.1, median 0 for GPs). The types of procedure reported in the data are described in Table 21.

• Investigations and tests are much more common than procedures for both PAs and GPs, although the largest single group remains those patients for whom no investigation or test was recorded (mean 0.36, median 0 for PAs; mean 0.34, median 0 for GPs). The types of investigation or test reported in the data are described in Table 22.

• The number of prescriptions from the index event was most commonly none or one for both groups (mean 0.89, median 1.00 for PAs; mean 0.78, median 1.00 for GPs).

• For referrals, a large majority of patients had no onward referral made at the index event (mean 0.9, median 0 for PAs; mean 0.11, median 0 for GPs) and the commonest type of referral was to secondary medical care. The types of referral reported in the data are described in Table 23.

• General advice is the most commonly recorded activity of the processes measured in this study, being reported for approximately half of the patient consultations, followed by advice on medication management and the use of over-the-counter medicines in a large minority (see Table 20).

Comparison of physician assistant and general practitioner consultations in recording process outcomes

The above descriptive accounts of the processes recorded in the consultations give some indication of where there may be similarity or difference between the actions of PAs and GPs.

Unadjusted analyses suggest no difference in the number of procedures carried out, the number of investigations or tests ordered or undertaken, or the issuing of Fitness to Work certificates, but suggest that patients consulting a PA are more likely to receive a prescription for medicine; PAs are more likely to make a referral to another service; and PAs are more likely to record giving general advice, advice on medication management and advice about over-the-counter medications (Table 24).

Once adjusted for covariates of relevance to each of the process outcome measures (detailed in Table 24 footnotes) and for clustering in the data at practice level, the unadjusted differences in the number of prescriptions, number of referrals and the giving of advice about over-the-counter medications diminish and are no longer statistically significant (see Table 24). Key in the adjustment of difference is the ‘PA study condition classification’ of the case as acute, chronic, minor/symptom, prevention or process. After statistical adjustment, PAs remain more likely to document giving general and medication management advice to patients (see Table 24).

Consultation process: length of consultation and advice taking

Length of consultation

The length of consultation recorded by the ‘patient enter’ and ‘patient leave’ functions of the practices’ electronic databases were available for 1812 (86.9%) consultations. Missing data were predominantly from three practices and for specific reasons: in one GP practice the surgery session management data (i.e. functional data other than the patient’s clinical record) had already been removed from the database before study data collection took place; in one GP practice the ‘patient enter’ time for telephone consultations was the time the patient telephoned to request an appointment, not the time they spoke with the GP; and in one PA practice the ‘patient enter’ time was infrequently used.

For the 1812 consultations with a time available, PA consultations were seen to have a significantly longer mean consultation time, 5.5 minutes in excess of the GP consultations (95% CI 4.8 to 6.2 minutes; p < 0.001 by t-test), in unadjusted comparisons (Table 25).

TABLE 25 - Length of consultation in minutes

Descriptive statistics	Consultation type
	PA consultations	GP consultations
n	896	916
Missing	36	238
Mean	16.79	11.28
Percentiles
25	11.00	7.00
50	15.00	10.00
75	21.00	14.00

Differences are seen by practice (Figure 4), although the lack of obvious pattern by newly qualified or experienced PAs, missing data about the length of experience in primary care of all the GPs and the variety of practice organisation of appointment times and type of work make adjustment for potentially confounding factors untenable.

FIGURE 4.

Box plot of length of consultation by practice.

In practice 11 (the furthest right in the box plot for PA practices), the mean length of consultation is obviously shorter than that in the remainder of the PA practices. Here, the PA undertakes both telephone and in-person consultations. Telephone consultations have a mean of 10.4 minutes (SD 8.9 minutes; range 3–41 minutes) and in-person consultations have a mean of 14.5 minutes (SD 5.7 minutes; range 4–47 minutes), being closer to (but still shorter than) the remaining PA practices. Similarly, in practice 8 (the furthest right in the box plot for GP practices), telephone consultations are also routinely carried out for same-day appointments, and the mean length of consultation for these is lower than that for in-practice consultations (although length of consultation data were available for only three such cases).

After adjustment for the ‘PA study condition classification’, age, age squared and the number of problems, the mean difference remained statistically significantly different for practice type (PA or GP) and for the interaction of practice type with age. This difference in length of consultation between PAs and GPs, however, varies by patient age, increasing with age before tailing off at the older ages. The marginal effects at different ages are shown in Table 26.

TABLE 26 - Marginal effects of length (minutes) of consultation at different patient ages

Patient age (years)	Difference in length (minutes) of consultation between GPs and PAs (marginal effects)
20	5.33
30	5.58
40	5.84
50	6.09
60	6.34
70	6.59
80	6.85

The average participant in this study has an age of 38.7 years, where there is a marginal (averaged over all covariates) effect of 5.8 minutes (95% CI –3.54 minutes to –8.06 minutes). All of these are significant differences.

Consultation outcome: reconsultation within 2 weeks

The study’s primary outcome is considered in terms of the numbers of patients reconsulting and the rate of reconsultation, a comparison of rates between PAs and GPs, the type of professional seen at the reconsultation and the planning of reconsultations.

The patient survey sample

The sample identified for inclusion in this element of the study consisted of 1020 adult (aged 16 years or over) patients (n = 490 PA consultations, n = 590 GP consultations). Figure 3 (see Chapter 4) details the identification of cases for inclusion/exclusion in the study, including the patient survey route of entry into the study. The survey sample differed from the overall sample of patients’ consultation records primarily by the exclusion of children aged under 16 years. Table 31 presents descriptive data of the survey sample against the overall consultation record sample for age, Table 32 for IMD, Table 33 for sex and Table 34 for ethnicity.

Response rate and respondents

Of the 1020 eligible patients given a survey, 52.8% (n = 539) responded with a completed survey, with 51.2% (n = 220) of those surveyed and consulting a PA responding and 54.1% (n = 319) of those surveyed and consulting a GP responding.

Tables 31 to 34 present descriptive data of the survey respondents against the survey sample for age, IMD, sex and ethnicity. These data suggest that survey recipients are older than the overall sample (explained by the exclusion of those aged under 16 from the sample) and that respondents are an older group still; that the IMD of survey recipients is higher than that of the overall sample but that of respondents is lower (indicating less deprivation among respondents); that sex is consistent across the consultation, survey and respondent samples; and that ethnicity appears to show the biggest difference between groups with survey recipients and survey respondents showing an increasing proportion of participants classifying their ethnicity as white British (although the reduction in missing data for these groups is dramatic and may have influenced the proportions).

The survey responses

The survey measured the patients’ experiences against a number of measures as follows: how good the PA or GP was at giving enough time, asking about symptoms, explaining tests and treatments, involving the patient in decisions about their care, treating the patient with care and concern, and taking problems seriously; confidence and trust in the PA or GP seen; satisfaction with the care received from the PA or GP at the index consultation; satisfaction with the general care at the general practice; willingness to recommend the general practice; and willingness to see a PA rather than a GP.

Table 35 describes the responses in total and by consulting with PA and GP groups and the unadjusted and adjusted comparison between PA and GP groups.

The majority of respondents were satisfied or very satisfied with their consultation with both PAs and GPs. The measures for which lower percentages of respondents reported themselves to be very satisfied were ‘explaining about tests and procedures’ and ‘involving you in decision-making’. There was no difference between PAs and GPs.

Positive correlation, significant at p < 0.01, was found between general satisfaction with the index event consultation and all other measures of satisfaction collected in the survey: giving enough time, asking about symptoms, listening, explaining tests and treatments, taking problems seriously, involving the patient in decision-making, treating the patient with care and concern, confidence and trust and willingness to see a PA. The majority of these measures were also correlated.

No statistically significant correlation was found between any measures of satisfaction and the study’s primary outcome, that is, the rate of reconsultation with the same problem.

Review of reconsultations for the same condition at the practice or urgent care facility

The assessment of the quality of computerised medical record entries for the key consultation components (using Weed’s SOAP classification101,102 as described in Chapter 2, The clinical review of reconsultations records), for cases with a reconsultation for the same problem (n = 388) at the practice or urgent care facility within 14 days, is presented in Table 37.

The PAs were judged to record the key consultation components more frequently in the computerised medical record, with each element reaching statistical significance for the difference. The PAs recorded appropriate index consultation records nearly 80% of the time (145 out of 183), whereas GPs made appropriate records in just under 50% of the time (99 out of 205; p < 0.001) (see Table 37).

This difference remained the same across all of the elements of the medical record that were observed. GPs had an excess of appropriate over inappropriate records only for investigations. The risk ratio of a PA making an appropriate record compared with a GP was 1.64 (95% CI 1.40 to 1.93; p < 0.001; see Table 37 for raw data), that is to say those consulting a PA are 64% more likely to get an appropriate medical record, as judged by the reviewers.

Review of reconsultations for a linked condition at the practice or urgent care facility

The assessment of the quality of computerised medical record entries as above for cases with a reconsultation for a linked condition (n = 75) at the practice or urgent care facility within 14 days is presented in Table 38.

The same judgement of the difference between the medical record entries of PAs and GPs for a linked problem as for the same problem was made: just over 90% of PA records (90.06%, 29 out of 32) and just under 60% of GP records (58.1%, 25 out of 43) were thought appropriate. Again, PAs were much more likely to make appropriate records, in all domains. The numbers are smaller, which may reflect in fewer significant differences between professions. The risk ratio of a PA making an appropriate record for these ‘linked’ consultations compared with a GP was 1.56 (95% CI 1.18 to 2.06; p < 0.001, see Table 38 for raw data), that is to say those consulting a PA are 56% more likely to get an appropriate medical record, as judged by the reviewers.

Review of all repeat consultations

The PAs made better records of their consultations, and the reviewers felt that they indicated safe practice. They did not feel that any of the consultation records represented unsafe practice.

Effectiveness of blinding/reviewers’ judgement of who conducted the consultation

The blinding of records to remove any indication of whether the consultation was with a PA or a GP was considered to be successful. Reviewers were unable to judge whether the index event had been consulted by a PA or GP in both the review of cases for the same condition and those for a linked condition.

For cases with a reconsultation for the same problem, reviewers gave an opinion in nearly all cases (98.2%, 381 out of 388); in the remainder of the consultations the reviewers stated that they were unable to decide. Where they could decide, there was low sensitivity and specificity; they judged 57% of PA consultations reviewed for the same problem to have been those of a GP (к = 0.196; Table 39).

For cases with a reconsultation for a linked condition, reviewers were similarly unable to judge whether the index event for these linked reconsultations had been carried out by a PA or GP, though they gave an opinion for all 75 cases. In the case of linked consultation, the reviewers judged 66% of PA consultations for a linked problem to have been those of a GP (к = 0.141; Table 40).

Overall, 58.3% (123 out of 211) of the consultations actually with a PA were thought on the basis of the records to have been with a GP. Thirty-nine per cent (57 out of 145) of the consultations that the reviewers thought were by a PA were actually by a GP. The raters were neither sensitive nor specific: diagnostic OR 2.36 (95% CI 1.58 to 3.53; Pearson chi-squared p < 0.001), sensitivity 42% (95% CI 35.1% to 48.3%) and specificity 77% (95% CI 71.4% to 82%).

Inter-rater reliability

Ten clinical review cases were rated by all four reviewers independently immediately following the review raters’ training exercise. The assessment of inter-rater reliability using Fleiss’s kappa103 for the assessment of the quality of computerised medical record entries for the key consultation components is presented in Table 41.

Taken against the consideration that a kappa score of 0.0–0.2 is ‘poor’, of 0.21–0.4 is ‘fair’, of 0.41–0.6 is ‘moderate’, of 0.61–0.8 is ‘good’ and of 0.81–1.0 is ‘very good’,104 the inter-rater reliability of the reviewers here was very good when assessing the quality of the computerised medical records, but very poor agreement on the type of practitioner (reviewers blinded to this) carrying out the consultation.

Comparison of reconsultation rates with judgements of appropriateness

The rate of reconsultation for PAs and GPs was presented in Chapter 5 (see Rate of reconsultion for the same problem), indicating that there were no differences in the rate of repeat consultations within the 2 weeks following the index consultation by type of professional seen at the index consultation.

After adjustment for the judgement of appropriateness (and practice clustering), the rate of reconsultation for the same problem at the practice or urgent care facility, or at the practice only, does not suggest any statistically significant difference between the rates for PAs or GPs:

• same problem at the practice or urgent care facility: rate ratio 1.124, 95% CI 0.918 to 1.374, p = 0.257 (see raw data in Table 42)

• same problem at the practice only: rate ratio 1.217, 95% CI 0.981 to 1.506, p = 0.074 (see raw data in Table 43).

Post-analysis observations of reviewers

At the post-review workshop, the GP raters, no longer blinded, felt that they had clearly mistaken the more thorough medical records made by the PAs for those of GPs. They wondered the extent to which this might reflect training: PAs might have more emphasis on systematic data recording; GPs might tend to make a more global assessment of the patient. Finally, they also speculated on the extent to which time to write records might be an issue (i.e. PA 15-minute consultation vs. GP 10-minute consultation).

Sample numbers: physician assistants and general practitioners who volunteered to be video recorded

A further 62 consultations were video recorded: 21 (33.9%) with four PAs and 41 with five GPs. The number of consultations varied between 3 and 13 for the GPs and their patients who consented to be video recorded, and between two and seven for PAs (Table 44). These consultations were then reviewed by at least two reviewers.

Leicester Assessment Package scores: how the physician assistant consultations compared with the general practitioner consultations

Grades were given by four raters using six of the seven main categories of the assessment package. These data were not normally distributed: median, interquartile range (IQR) and Mann–Whitney U-test to compare GPs and PAs are presented in Table 45. Higher median scores were found for GPs than for PAs across all the dimensions of competence in the LAP. GPs were generally rated above 50% for each element of the LAP; PAs’ median scores were mainly 60%, with some at 40%, although the lower range score was the same for GPs and PAs on all competence dimensions bar problem solving. It was noted by the assessors that the PAs gave the computer more attention and may be writing better medical records.

The weighted scores showed similar differences; GPs scored higher across all competencies (Figure 5). The median overall percentage for the practising GPs was 58.6% (25th centile 50.7%; 75th centile 69.7%). For PAs, the median percentage was 47% (25th centile 30.6%; 75th centile 60.0%). These data were not normally distributed; the differences were statistically significant [Mann–Whitney U-test (two-tailed) p = 0.012].

FIGURE 5.

Range of results with the LAP: a box plot comparing PAs and GPs.

The raters did not consider any of the health-care professionals to be unsafe to practise.

Inter-rater reliability

The Cohen and Fleiss kappa,103 also referred to as ‘Cohen’s kappa’, was used to measure inter-rater reliability. Across all the ratings, kappa just fell into the poor agreement category104 (к = 0.193); there was fair agreement104 for the GP assessments (к = 0.225), but poor agreement104 for the PA assessments (к = 0.121).

Workshop with raters

At the workshop, the GP raters, no longer blinded, felt that the following themes should be taken into account when comparing records:

• PAs were much more systematic in their approach to record writing, and GPs were much more likely to make global assessments in their records. Possibly there were differences in training with regard to records and what to record when assessing patients.

• Raters felt that there was a difference in caseload between GPs and PAs, with the GPs having a more complex caseload. However, the PAs were effective in managing a core part of the primary care caseload, and the nature of emergent histories was such that many of the cases were complex.

• Raters felt that the longer PA consultations may allow the creation of better medical records.

• The only aspect of PA consultations that raters would criticise was that some PAs did not correct the patient calling them doctor. Clearly, the patient was comfortable with this misapprehension; however, this appeared undesirable and reminiscent of when medical students were introduced as ‘young doctors’.

• The need to get prescriptions signed appeared to fit into the workflow in the consultation, much as in the way that GPs support nurse prescribing.

• Raters reiterated their view that they did not identify any patient safety problems, and that PAs demonstrated expertise across a range of consultations.

Understanding and conceptualising the role of the physician assistant

The participants’ accounts of consultations with PAs revealed diversity in terms of being familiar with the PA who they consulted in this study. A small number of participants confirmed that this had been their first contact, while others had had varying levels of previous contact; for example:

I’ve actually seen [PA name] in the past, I had earache sort of, I don’t know, some time last year and I’ve actually seen him before.

Participant 15

Some participants also described prior interaction owing to the PA’s previous treatment of a spouse or child. Three participants described an ongoing relationship with the PA as their clinician in the treatment of a long-term condition.

It was in these descriptions of the relationship with the PA that some of the participants revealed how they understood and conceptualised the role of the PA, with many using the word ‘doctor’ to refer to the PA, as in this example:

Mr X, do you remember is this the first time that you’ve seen this particular health professional?

By health professional you mean my GP or hospital visits?

The medical, no, the medical care practitioner at your GP surgery.

Err, Dr X [the PA]?

It was evident that there were accurate and inaccurate conceptualisations of the PA role.

Those with an accurate conceptualisation of the role appeared to derive this from previous personal or family experience and/or because the practice routinely informed its registered population about the PA role with leaflets, posters and explanations. These participants conceptualised the PA role in relation to doctors, a ‘GP with a difference’ (Participant 22), or, as in this more detailed illustration:

They’re just like doctors . . . I asked them about physician assistants and they gave me a leaflet and then I had a chat with the person . . . and he explained about his education, and his background . . . and you know I realised then that they’re almost doctors, they just can’t do everything here that they can, that a doctor could do.

Participant 17

The inaccurate conceptualisations reported included believing that the PA was a professional in training, a nurse, a qualified doctor from another country who could not prescribe in the UK or a qualified medical doctor. Unfortunately, some in this group had not known that the person they had seen was not a doctor until this information emerged during the research interview, and were puzzled as to why they were not given more information, as in this exemplar:

I called him doctor because I thought he was a doctor . . . I don’t know why they didn’t tell me, I’m not sure whether they didn’t want me to think he wasn’t a doctor and to think that he wasn’t going to do such a good job, but I was very open to whatever he was going to say to me anyway so I wasn’t judging him because he wasn’t.

Participant 03

These participants reported a varying level of both being informed and being able to recall the information about this new role. There were also those who described receiving information from the practice or the PA but then not understanding the implications of that information, described as in this exemplar:

She did explain, she said I’m from America.

Participant 31

Several participants described a deductive process whereby they had picked up on clues that the PA was not a doctor, and ‘worked that out myself’ (Participant 29), most obviously with the issue of prescription signing, as in the next example:

Then when I went to her to write out a prescription, she said, ‘No, I must get the doctor to sign it.’ Then I realised. So, I think that’s the right way to go about it because it’s better because then if the patients are, ‘Oh no, she’s not a doctor . . .’

Participant 32

Despite the inaccuracies in conceptualisation and the lack of information for some, most of these participants then turned to describing positive qualities or a desirable outcome from seeing the PA. This mirrored the positive views expressed by those who had an accurate conceptualisation of the role and links to the next theme of trust and confidence in the PA.

Trust and confidence in the physician assistant

The issue of trust and confidence in the PA was raised in most of the interviews, in response either to general questions about the consultation or direct questioning by the interviewer. Where the response was to a direct question, participants generally used unambiguously positive words such as ‘certainly’ or ‘absolutely’ or gave a percentage score, for example:

Did you feel trust and confidence in that episode of care?

Definitely. Yeah, 95–99%.

The reasons for the levels of trust and confidence can be grouped into subthemes: the personal attributes of the PA, demonstration of clinical competence and positive outcomes of the consultation, similarity of the consultation with that with a GP, and trust in their GPs and the wider NHS to provide appropriate, safe and effective care. Each of these is discussed in turn. The section ends by considering explanatory factors given by patients who did not express high levels of trust or confidence.

The impact of physician assistants’ lack of authority to prescribe medications

Participants reported having become aware that PAs were unable to prescribe either through the explanations given to them prior to the appointment or during their consultation. For those participants who described having required a prescription, different methods to organise this had been experienced. Most reported the PA leaving the consulting room to discuss the case with a GP and then return with the signed prescription while the patient waited either in the consulting room or in the waiting room. Some patients also reported collecting the signed prescription from reception then or later, or having it faxed to the pharmacy.

The need for prescriptions to be verified and signed by a GP was reported by most participants to cause no apparent or significant delay. A small number of participants reported delays of 5–10 minutes, with a minority reporting longer waiting times ranging from 15 to 30 minutes. Short delays appeared generally to cause no particular problem to participants and were considered reasonable or understandable. A minority, however, felt that any delay caused by the PAs’ lack of prescribing authority was problematic, as exemplified here:

It’s quite annoying, actually, because, I mean, I feel that if people can prescribe it they should be able to sign it.

Interview 23

Participants, however, were mostly understanding of this potential limitation and it appeared that a consultation experience that was in other ways a positive one was not viewed negatively because of the need to wait for a prescription.

Each of the above themes features in responses to the final issue explored, which was participants’ willingness to consult a PA again.

Willingness to see a physician assistant again

A small number of participants had actively sought an appointment with a PA and gave the following types of reasons: a shorter waiting time to see a PA than a GP, dissatisfaction from prior appointments with GPs and trust in the PA based on previous contact. Most of participants reported that they had not been offered a choice of whether they saw a PA or GP when they booked their same-day appointment. Those who had been offered a choice chose the PA, as this option provided them with an earlier appointment.

Most of the participants stated a willingness to see a PA again for any condition, while one participant stated that they would not wish to consult a PA again as it had resulted in delays in a consultation with a GP.

The willingness of other participants to consult was expressed as contingent on a number of factors. Some participants who were interviewed at a point very close to the time of their consultation with the PA described their willingness to consult as contingent upon the success of their current course of treatment. Others described themselves as willing but differentiated which professional was appropriate to consult according to how serious or complex the problem(s) were that they wished to consult about. Broadly speaking, participants with complex medical conditions or more complex prescribing needs considered it more appropriate and preferred to consult a GP rather than a PA. Minor conditions or complaints were seen as appropriate for consultation with a PA, as in this example:

I think if it was just a general complaint he [relative] wouldn’t mind seeing [the PA] but regarding his prescription, he’s a bit fussy about his medicine, he would prefer to see a doctor.

Participant 02

Willingness to return to see a PA again also appeared to be influenced by the participants’ beliefs about pressures faced by the GP and recognition that not all medical problems required the attention of a doctor, as in this example:

I am very aware that I don’t want to take up appointments when it isn’t really that necessary. So the thought that there is a role within the surgery where I could go and see somebody who isn’t as pressurised as the doctor, to see, to understand, to and go and see them for something that’s like not trivial particularly, but I don’t necessarily need to see a doctor for then I feel that that role is a really good thing to have in the surgery and I feel that I would be happy to utilise that again, definitely.

Participant 03

For some participants, regardless of how satisfied they were with the PA (or the GP) in a single consultation, maintaining continuity of their care with one professional was considered equally as important as, if not more important than, having a preferred type of practitioner. Consequently, participants who had already consulted a GP or a PA about a particular problem preferred to consult the same practitioner again, to achieve continuity of care:

I’m trying to make her my regular, I say doctor, but my regular person I see at the surgery . . . she seems to understand my needs. I get on really well with her. She listens to me; she makes me feel very comfortable. She seems to, you know, she gives good advice as well. She understands my situation very well, and she’s the only one, she’s actually the only one at the doctors at the moment that seems to be able to give me that response. As I said, because it’s a [type of practice], you’re sort of shuttled around from doctor to doctor. You don’t really get to make a relationship with anybody.

Participant 21

Other participants also referred to the benefits of their longer-term relationships with the PA and cited this as a reason for willingness to reconsult with a PA. Some gave positive accounts of the PA’s ability to recall personal details such as a medical history, the names of extended family members, or work and social circumstances, as well as seeing the PA as part of the community.

The patients’ experiences

The majority of patients who responded to the survey in this study reported that they were satisfied or very satisfied with their consultation with the PA (see Chapter 5, Patient survey). These findings mirror the national results of general practice patient surveys. 197 A high level of patient satisfaction with general practice consultations is not an unusual finding; however, this is the first time that patient satisfaction concerning PAs in the UK has been reported from such a large sample drawn from a variety of general practice settings. This mirrors the findings in medium- and large-scale surveys of primary care patients consulting PAs in the USA. 155,160,168–171

Patients consulting a PA described factors such as continuity of clinician and perceived complexity of their illness as important in their willingness to consult the PA again or their preference for consultation with a GP. These types of factors have been evident in other studies regarding patient views on consultations in primary care. 198,199 A small number of patients indicated their preference to always consult a GP in future consultations (see Chapter 5, Patient survey). This is important in that studies investigating patient preferences following consultation with another type of clinical professional (nurse or pharmacist) in UK primary care have reported one-third or more expressing a preference for future consultations with a GP despite high levels of satisfaction with the initial consultation. 90,92,200,201

It is noteworthy that, of the patients who were surveyed in practices not employing PAs, 28% were willing or very willing to consult a PA and 20% were neither willing nor unwilling (see Chapter 5, The survey responses). This is a lower percentage than in studies carried out in three other countries where PAs are relatively unknown. Those studies, however, were more in-depth, hypothetical studies posing scenarios of different lengths of waiting time to see the PA or a physician in ambulatory care settings. 173,202,203 One explanation in the UK setting may be that a common feature of general practice since the 1990s has been the employment of other types of staff such as nurses,20 which is not as common in many other countries.

The average length of consultation for same-day appointments with a PA (16.79 minutes) was significantly longer than the average consultation time with a GP (11.28 minutes). The national workload survey has reported 11.7 minutes, 12.1 minutes, 16.4 minutes and 15.5 minutes as the mean consultation lengths with a GP partner, salaried GP, GP registrar and NP, respectively,204 suggesting that the study’s data for GPs, at least, are representative. Previous UK studies of PAs do not report length of consultation. 60–63 Past UK studies of nurses providing this mid-level practitioner activity with same-day-appointment patients also report longer consultation times for GPs. Kinnersley et al. 92 reported a median of 6 minutes (range 4–8 minutes) for GPs and a median of 10 minutes (range 7–14 minutes) for NPs, reducing to 8 minutes (range 6–11 minutes) when the time for obtaining a signature on a prescription was removed. Venning et al. 91 reported a mean consultation duration of 11.57 minutes for NPs (SD 5.79 minutes) and 7.28 minutes (SD 4.80 minutes) for GPs. Shum et al. ,90 in a study of practice nurses with additional minor illness training providing same-day appointments compared with GPs, reported that both were given 10-minute appointment slots and that the nurses had a mean consultation time of 10.2 minutes (SD 5.7 minutes) compared with a GP consultation time mean of 8.3 minutes (SD 6.4 minutes). Further investigation is required as to the variation in consultation lengths by differently skilled and experienced clinicians adjusted for patient case mix and when using surgery systems such as telephone triage.

Patients reported varying levels of understanding and conceptualisation of this new professional role in UK primary care. Similar findings were reported in studies investigating understandings of general householders in the early introduction of PAs in the USA. 172 The practices in the current study had a range of strategies to inform their patients about the role, some more evident than others. Our interviews suggest that those practices using multiple methods, including a script for receptionists and explanation by the PA, were more successful in achieving understanding. This suggests implications for operational practice in general practice and a role for the national PA organisations in sharing best practice.

The patients’ outcomes

All of the patients in this sample were attending ‘same-day appointments’, that is having a problem or symptom that required consultation that day as decided by the patient or the practice using telephone triage. GPs were consulted by an older patient group with a higher incidence of measures of comorbidity than those patients consulting the PA (see Chapter 5, Consultation records). PAs tended to be consulted by more patients with problems classified by the study as ‘minor/symptoms’, while GPs were consulted by patients whose problems were more likely to be classified in this study as ‘acute’ or ‘chronic’, that is to say having greater medical acuity or complexity in terms of comorbidity (see Chapter 5, Characterisation of the health problems of the consultation record sample). Previous UK and international studies of PAs in primary care have not included a measure of severity or acuity of presenting medical problems. One recent US study also has evidence that PAs see patients who are less medically complex than those seen by doctors,193 using the patients’ diagnostic care group (DCG) score as the measure of complexity. DCG is a predictor of future health-care expenditure based on age, sex and comorbidities. 205

After adjusting the data for covariates and clustering, it was found that consultations with PAs resulted in the same rate of reconsultation for the same problem in the 2 weeks after the index event as with GPs: our primary outcome measure. They also resulted in the same rate of process outcomes: procedures, investigations or tests ordered or undertaken, prescriptions, referrals and advice on over-the-counter medicines (see Chapter 5, Consultation records). This is the first time that these process and outcome measures have been reported for PAs in the UK primary care setting. One US study from the 1970s showed that two PAs in comparison with nine family practice physicians had higher medication costs, although they also had higher laboratory costs (not found in the present study) and higher rates of good outcomes for patients. 167 A more recent study of a single PA in the Netherlands showed no difference in referral rates to secondary care from the PA compared with those of the GPs. 164 The only other UK evidence of mid-level practitioners, from studies comparing NPs with GPs in same-day consultations in the UK, provides more mixed evidence. Similar to our findings for PAs, Kinnersley et al. 92 reported no difference between NPs and GPs in the issuing of prescriptions and making referrals. Shum et al. 90 also reported that nurses and doctors issued prescriptions for similar proportions of patients. In contrast, Venning et al. 91 reported that NPs undertook significantly more procedures, ordered more tests and investigations and were significantly more likely to ask the patient to return within 2 weeks.

The adjusted figures also showed that PAs were more likely than GPs to have documented that the patient had been given general and medicine management advice (see Chapter 5, Comparison of physician assistant and general practitioner consultations in recording process outcomes). It could be argued that this may be an artefact of differences in record-keeping styles between doctors and PAs. It could equally be associated with the longer consultation time of PAs. The variation in electronic record keeping in general practice has been noted before206 and the more detailed completion of records by NPs has been observed in comparison with GPs in the UK. 207 In the current study, however, differences between professional groups were not apparent to the independent GP assessors, who were blinded to whether the author of records was a GP or PA (see Chapter 6). This finding has not been reported before in the UK or international settings. For another mid-level practitioner group, NPs, a UK study also found that they delivered more advice in longer consultations. 92 The analysis of the audiotaped consultations with NPs and GPs reported that nurses talked significantly more about treatments, including how to carry out or apply a treatment. 208

Impact on the organisation of the practice

From the scoping survey (see Chapter 3, The scoping survey), it was evident that those closely associated with general practice could see a potential role for PAs within general practice teams but were looking for more evidence with regard to impact. From the case study element (see Chapter 4), the PA and practice staff interviews revealed that PAs were employed to fit into the organisation of the practice and the aim was for minimal impact per se, as with any other member of staff. The newness of both the occupational group and, in some instances, the individual’s qualification had some effect. For those practices new to employing PAs, one effect was the need to ensure that other staff understood this new-to-the-UK role. The next effect was a need to assign a supervising doctor and ensure time for that supervision (whether about an individual patient consultation or as a regular supervision session), and for newly qualified PAs to arrange ongoing professional development. The extent to which the role of the PA developed within each practice appeared to depend both on the interests and knowledge of the PA and on the opportunities presented within the organisation of the practice. This is captured in the discussion of deployment, earlier in this chapter (see The deployment of physician assistants in general practice).

Impact on the working practices and relationships with other professionals

In the case studies, it was evident that the PAs had an impact on the working practices of some of the GPs. In some practices, they were employed specifically to release GP time for attending to more complex patients or other aspects of the GP workload. GPs described how more of their work was with patients with more complex needs and patients outside the scope of practice of the PA (see Chapter 4). This was confirmed in the clinical consultations entered into the case study, where it was evident that PAs were consulted by younger patients with a lower incidence of measures of comorbidity compared with those consulting GPs (see Chapter 5, Consultation records). A previous UK study of a smaller sample of PAs suggested that PAs saw older patients than GPs but concurred that the PAs saw a similar but less complex case mix to GPs. 62 This was also evident in the observation of clinical meetings (see Chapter 4), in which the GP partners were focused on the care of patients with multiple physical, mental health and social problems and also on ensuring that salaried GPs, GP registrars and nurses were supported in the management of such patients.

The extent to which PAs impacted on the work in each surgery of the supervising doctor varied depending on the experience of the PA and the systems in place to signal the PA’s need for consultation about a patient and/or prescription signing. While PAs’ lack of authority to sign a prescription was an issue in all of the practices, it should be set in the context that, firstly, in most of the practices the nurses did not have prescribing qualifications and were also dependent on efficient systems (see Chapter 4) and, secondly, that most consultations (with either a GP or a PA) did not result in the issuing of a prescription (see Chapter 5). The types of actions to minimise the consequences from lack of authority to sign prescriptions have been noted before in the UK context61–63,72 and in other countries developing the PA role. 164,209

The boundaries between the work of the PA and that of the nurses in the practices were described as delineated and agreed without overlap (see Chapter 4), a point also noted in the primary care pilots of PAs in Scotland. 61 The PAs were reported to be flexible in their skills, and so were often able to cover nurse absences, a feature again noted in the earlier study of PAs in England. 60 It was evident from the interviews with the practice managers and GPs (see Chapter 4) that the priority was to deploy available staff most efficiently against the practice service delivery needs and the demands of each day. While some GPs noted that there had been disquiet from some NPs when PAs were first employed, we were able to find evidence of only positive relationships and working practices between PAs and nurses in the case studies. This has been noted before in the UK60,61 and in other settings where PAs have been introduced in primary care more recently. 209

The impact of including the physician assistants on practice costs

This study has shown that, after controlling for case-mix differences, GP and PA consultations resulted in similar levels of prescribing, referral, tests, procedures, reconsultations and patient satisfaction. The observed difference in consultation times between PAs and GPs (16.8 minutes vs. 11.3 minutes) is very similar to the national data for NPs and GPs (15 minutes vs. 11.7 minutes). 119 After adjusting for covariates, a PA consultation for the ‘average’ patient is 5.8 minutes longer than a GP consultation for the same patient, at a marginal cost of £6.22. Although PAs’ consultations are longer than GPs’, they document giving more advice to their patients, and the lower costs associated with their consultations mean that they still deliver care more cheaply. This analysis does not, however, take account of extra costs associated with the use of PAs, including GP time in supervising, training and signing prescriptions, which was not quantified in this study.

In the absence of other studies of the relative costs and effectiveness of PAs and GPs in the UK setting, comparative evidence can be drawn from research that has compared GPs with NPs, another mid-level practitioner group. The findings of one randomised controlled trial in English general practice over a decade ago were similar to those in the current study of PAs; the NPs had similar rates of prescribing, referring, testing and reattending as GPs, and they delivered more advice in longer consultations. 92 Another UK-based comparison of GPs and NPs at about the same time concluded that the clinical care and costs of GPs and NPs were similar; a higher propensity of NPs to order tests and for patients to return for follow-up appointments offset lower consultation costs. 91 A systematic review showed that patients were more satisfied with NPs than with doctors, and while there were no differences in prescribing, referrals and reconsultations, NPs had a higher propensity to order investigations. 39 A modelling study based on the review calculated that there was no significant difference in costs from the NHS or practice perspective because the GP time that goes into the supervision and reconsultations arising from the use of NPs offsets the lower consultation cost per minute of NPs, although sensitivity analysis showed that if the supervisory time input of GPs was halved, NP consultations were cheaper. 210 In a recent study in the Netherlands, NPs provided equivalent outcomes to doctors for a lower cost; the mean saving per consultation was € 8.21,211,212 which is similar to the saving recorded for PAs in this study of £6.22.

Further investigation

This study has highlighted a number of questions that require further investigation. These include:

• the contribution of PAs in other first-contact services as well as secondary services

• the contribution and impact of all mid-level practitioners, PAs and NPs in first-contact services

• the extent and impact of supervision and conferring on patients for mid-level practitioners, PAs and NPs in different types of primary care settings, for example GP practices or urgent care facilities

• the influence of sociodemographic and geographic factors, such as inner-city, remote and rural settings, on the demand for PAs and other mid-level practitioners, in a flexible primary care workforce

• the reliability and validity of classification systems for both primary care patients and their presenting condition and their consequences for health-resource utilisation

• the factors and influences in GP partner and manager decision-making as to staffing and skill mix in general practice, with particular reference to beliefs, preference, evidence and use of financial data.