Investigating the contribution of Physician Assistants to primary care in England

This proposal addresses the question of the changing workforce in primary care within the research brief. A number of new work roles are being developed in health care in the UK. People undertaking these roles often work across the professional boundaries of more established roles such as those of the nurse or doctor. Many of these roles are collectively termed 'mid-level' practitioners, including one called the Physician Assistant (PA).

The PA role was established in the USA in the 1960s due to a shortage of medical doctors, particularly in rural areas. Preparation to become a PA consists of a two year university-based medical education and clinical practice experience. Of the almost 63,000 PAs practising in the USA, 26% are in general practice, as opposed to 12% of medical doctors. PAs examine, investigate, diagnose and treat patients to a level agreed by the doctor who supervises them. Research evidence from the USA suggests that they deliver care that is as safe and acceptable to patients as that of doctors, but with less efficiency, and cost savings vary. In England there are estimated to be approximately 40 PAs in practice.

Two studies of pilot projects employing USA-trained PAs in the UK reported that PAs consulted GPs infrequently about the patient's treatment plan and showed greater flexibility and diagnostic skills than extended roles in nursing. However, PAs were limited by not being able to prescribe; and higher than anticipated costs were encountered. Four UK education programmes for the PA role have commenced in recent years, with very small numbers of UK trained PAs in practice to date, although this number is set to grow. Primary care services are currently being developed, increasing the amount of care provided outside of hospitals. A variety of skills and roles is expected to be needed.

The outcomes being measured for PAs in general practice in this research are utilisation; conditions and characteristics of patients; patient choice; patient experience; consulting style; rates of investigations, prescriptions, referral to other members of the practice team and to specialists, and re-attendance for the same complaint; cost; and factors supporting or inhibiting the employment of PAs.

The research will be carried out in two main parts, and using a mix of methods. The first part will look at the research outcomes at an overview level, using three methods: a review of published research and unpublished materials on PAs to update the literature known to the applicants; interviews with up to 40 key informants from central departments of health, regulatory organisations, patient organisations and those who commission services; and an electronic survey of how PAs currently work in general practice, sent to known employers of PAs. The second part will look at the research outcomes in detail using six general practice 'case studies'.

Three will employ at least one PA but each practice will be different in terms of practice size and patient characteristics, staffing, geography, and the types and roles of the PA/s employed. Three practices will not employ a PA, but will be matched to the characteristics of the three 'PA practices'. Data to study these cases in depth will be gathered in seven activities: 1) interviews (up to 30) with PAs, GPs, other practice staff and local representatives with a patient, commissioning or secondary care perspective; 2) Collection of anonymised electronic patient records (for two months) to describe the types of patients and conditions seen, care provided, outcomes of consultations (prescriptions, investigation, and referral), and re-attendance for the same problem; 3) patient surveys for two months of consultations with a PA, GP or nurse; 4) patient interviews (up to 45 consulting a PA); 5) work activity diaries for PAs, GPs and nurses for two 1 week periods (summer and winter); 6) video recording at least 10 PA and GP patient consultations; 7) collection of practice papers or reports on clinical performance, workload, and funding.

Data from all seven activities will be drawn upon to describe the case studies in detail. Analysis of any qualitative data (that is, verbal descriptions or images) will be thematic. Analysis of quantitative data (that is, numerical data) will involve statistical description, and comparison of different case studies and staff groups within case studies.

The ethical issues involved in this study are those of consent to take part in any of the activities. NHS ethical review and approval and local research governance approval will be sought. Individual members of staff and patients who will be participants will be asked to consent individually. In the case of the video recordings, patients will be asked to consent both before and after the consultation. All data will be pseudonymised (that is, an identity number will be ascribed, with all identifiable data removed from the research records).

This team is well placed to carry out the research due to its track record in carrying out research in primary care, new roles and patient choice and patient experience. The team brings a range of expertise and perspectives, health services research, health economics, bioinformatics, service user involvement and sociology, as well as prior and current primary health care practice experience.

The costs requested are mainly for staff time to oversee and carry out the research, plus contributions of the applicant team to study direction, project management, data collection, analysis and report writing; and a full time researcher for the duration of the project. Service user involvement will also require resource to offer reimbursement and expenses in line with national guidelines to those who contribute time and expertise.

Some additional monies are required to securely store the video recordings of confidential medical consultations. Additional minor costs are requested for administrative items (including postage). NHS costs will also be requested after the first part of the study is completed for the NHS staff time required as participants, and in accessing data