Journals Library

Understanding medical and non-medical antibiotic prescribing for RTIs in primary care out of hours services - the UNITE study

Project title

Understanding medical and non-medical antibiotic prescribing for RTIs in primary care out of hours services (The UNITE Study)

Project reference


Final report date

17 October 2016

Project start date

01 October 2015

Project end date

30 September 2016

Project duration

12 months

Project keywords

Respiratory tract infections, Infections, Antibiotic prescribing, Qualitative, Out of hours, Primary care

Lead investigator(s)
  • Dr Geraldine Leydon, Accademic Unit of Primary care and Population Science, University of Southampton
NIHR School Collaborators
  • Samantha Williams, Senior Research Assistant, Primary Care and Population Sciences, University of Southampton
  • Professor Michael Moore, Professor of Primary Health Care Research, Department of Medicine, University of Southampton
  • Professor Paul Little, Professor of Primary Care Research, Department of Medicine, University of Southampton
  • Professor Sue Latter, Professor of Nursing, University of Southampton
  • Dr Sarah Tonkin-Crine, Senior Researcher and Health Psychologist, Nuffield Department of Primary Care Health Sciences, University of Oxford
  • Dr Caroline Eyles, Senior Research Fellow, Department of Medicine, University of Southampton
  • Dr Amy Halls, Qualitative Research Fellow, Department of Medicine, University of Southampton
  • Dr Lisa McDermott (Kings College London)

  • Dr Karen Postle (PPI) 


Project objectives

In light of the increase in antibiotic prescribing for Respiratory Tract Infections (RTI) in primary care OOH services the study aims to explore medical and non-medical prescriber’s views on and experiences of prescribing antibiotics for RTI in primary care OOH services.

  • OBJECTIVE 1: Identify GP and NP experiences of prescribing antibiotics for RTI OOH and explore facilitators and barriers to reducing antibiotic prescribing.
  • OBJECTIVE 2: Explore similarities and differences between GP and NP antibiotic prescribing and explore their views on the need for a training intervention and views on content and mode of delivery of such an intervention.
  • OBJECTIVE 3: Compare OOH findings with existing evidence on in hour antibiotic prescribing.
  • OBJECTIVE 4: Feedback findings to key stakeholders to determine the need for a targeted OOH intervention to optimise prescribing behaviours.

Changes to the project originally outlined 

Due to a similar study being run by PHE, we experienced a 5-week delay at the start of the study. As subsequent meetings and communications were required, we were unable to start the study at the desired time. This caused a delay of 5 weeks, but also meant the study team needed to spend time altering study paperwork in order to avoid overlap with the PHE TAP Project.

Recruitment therefore began on the 1st December (a 9-week delay from original planned start date), which in turn was a difficult time to recruit due to its proximity to Christmas. We therefore requested a 10-week extension to allow for this delay.

Brief summary


Three different methods were used to identify potential participants for interview using the Clinical Research Network (CRN), the Health and Social Care Information Centre and the Association for Prescribers. All invitees were sent a participant information sheet, a reply slip and a consent form. Informed consent was received before any research activity took place. All responses were sent to the lead researcher (SW) who then contacted the participant to check eligibility and consent. Once consent and eligibility were confirmed, a convenient date and time for a telephone interview was arranged. Participants taking part were each paid an honorarium of £50. Interviews were conducted between November 2015 and April 2016.


GPs and NPs working within primary care OOH services were recruited and interviewed over the telephone. NHS OOH Services were defined as those that operate outside of normal working hours (6pm to 8am) that can be accessed by calling NHS 111 and or walk-in patients. To ensure diversity, purposive maximum variation sampling was used to identify key informants (Patton, 1990). Snowball or chain sampling, a technique for locating further participants through existing ones (Patton, 2002), supported this sampling strategy. The primary sampling criteria was to ensure an equal mix of GPs and NPs that work in rural and urban settings as well as staff from different OOH organisations. Secondary criteria sought to explore whether prescribers worked solely in OOH or a mixture of in-hours practice and OOH, and the number of years clinical experience to monitor for mix during recruitment. Prescribers were excluded if individuals solely worked in Secondary Care OOH to permit sole exploration of OOH services in primary care settings. Practitioners who were not available to undertake a qualitative telephone interview or provide informed consent were also not included.


A semi-structured interview guide was developed (see supplement) from the literature and built on questions from a topic guide exploring GP views in relation to delayed antibiotic prescribing for RTIs (Ryves et al, 2016). Interviews were semi-structured and focused on antibiotic prescribing for RTIs in primary care OOH services, the use of strategies to reduce unnecessary antibiotic prescribing, such as delayed prescribing, the influence of patient expectations on prescribing decisions and training in antibiotic management for healthcare professionals. The interviews were conducted by two experienced qualitative researchers (SW and AH) and were audio recorded an transcribed verbatim to facilitate analysis.


Whilst interviews were taking place the team met regularly with SW to discuss topics raised by participants, and the interview guide was refined as the interviews continued. Interview transcripts were read and reread and audio recordings listened to by SW both during and after the data collection period. An inductive thematic analysis approach (Braun & Clarke, 2006) was used to analyse data drawing on methods of constant comparison (Walker &

Myrick, 2006). SW independently coded the first five interviews and SW, STC, AH, KP and CE met to agree on preliminary codes. Following continued coding of a further five transcripts, SW developed a draft coding frame that was discussed and agreed by the team. STC, AH and KP independently coded six transcripts (20%) using the coding frame; discrepancies were minor and strategies for resolution were implemented. Codes were then clustered into higher-level descriptive themes, where both within- and between-participant contradictions and inconsistencies were considered. Theme labelling and interpretation was continually discussed in regular team data meetings. Saturation was reached for the major themes. NVivo 11 for Windows was used to facilitate the analytic process. ID numbers have been used throughout to maintain participant confidentiality.

The core team had a diverse range of backgrounds: SW is an allied health professional and research fellow; KP is a PPI representative with past experience in qualitative research methods, GL is a sociologist and associate professor with expertise in qualitative methodology; STC is a research health psychologist with expertise on antibiotic prescribing; AH is a sociologist and a qualitative research fellow. Our epistemological position may be best characterised by subtle realism (Hammersley, 1992). Our goal was for a pragmatic analysis; increasing understanding of this area from a prescriber perspective and leading to suggestions for future research and practice.


A total of 1253 prescribers were approached to take part in the study. 112 (8.94%) individuals responded to express an interest to participate in a telephone interview. See Consort Diagram 1 for recruitment details. 15 GPs and 15 NPs took part in qualitative interviews. The final 30 interviews were arranged with participants who completed study paperwork. The level of interest in the study allowed the team to purposefully select participants. Illustrative quotations are not provided within this report, but will be included in a published manuscript.

Theme One: Managing RTIs in Primary Care OOHs

Participants described the process of communicating decisions about treatment and how they explain their decisions to patients when they do not require antibiotics. Three stages were described which were reported to be iterative in nature: negotiating treatment, managing patient expectations and safety netting. Participants described factors that both facilitated and hindered this process, and highlighted how these factors influenced their ability to move through the explanation and decision making process with their patients.

Theme Two: Factors influencing the decision to prescribe antibiotics in OOHs (Objective 1)

Organisational and individual factors, including both prescriber and patient factors, were reported to influence antibiotic prescribing decisions. Organisational factors that reportedly influenced antibiotic prescribing in OOH included limited access to patient records, consultation time, working contracts and the amount of audit and/or feedback were reported to be important in OOH prescribing.

Prescribers reported that accountability and prescriber responsibility influenced their decision-making, although this varied between GPs and NPs (Objective 2). NPs reported being more accountable for their prescribing than their GP counterparts. Clinical experience and training were both reported to influence decision-making and again this varied between GPs and NPs. GPs suggested that NPs would be more cautious in their prescribing and therefore protocol driven, whereas GPs reported being guided by ‘gut feeling’ and clinical intuition.  

Clinical interviewees described the patient population seen in OOH as distinct from those seen in general practice within usual hours. Participants described the particular pressure to enable patients to return to work quickly. Participants reported a perception that patients are generally sicker when they attend OOH services. With a lack of patient follow up, this appeared to translate into a perceived higher risk of negative clinical outcomes clinical workload and therefore a much higher likelihood of antibiotics being prescribed.

Non-clinical patient factors were also reported to influence prescribing decisions including 

perceived patient anxiety, a patient’s ability to re-access OOH, patient-practitioner rapport and assessing patients to determine their ‘sensibility’ and understanding of the topic discussed within a consultation.

Theme Three: Antibiotic prescribing training for staff in OOHs

The third and final theme from described training for antibiotic prescribing in primary care OOH services. Participants reported barriers and facilitators to the process, what they would like to see in terms of delivery and main provider, and the skills they would like to be incorporated should training be provided on this topic.

Participants reported three main barriers to providing training in OOHs, and these are time, finances and staffing issues. Participants highlighted that there were ways to encourage individuals to take up training such as providing Clinical Professional Development (CPD) accreditation, ensuring the training was a mandatory requirement provided this was part of a clinical update and allowing protected time in an environment conducive to learning.

Clinicians reported that they would like to see training made available in different formats, including online and face-to-face with a simple structure. It was thought training should be delivered by a skilled professional still working in clinical practice, and hosted by their main employer or CCG. Most reported they would be happy to have training with varied audiences of mixed professional groups.  

GPs and NPs reported they would like updates on the following topics: Patient management and self-management tools and or advice, and consultation and communication skills. Most importantly, participants described how training should incorporate the most up to date guidelines on treating RTIs and information on accessing and utilising tools when not prescribing an antibiotic, such as the RCGP toolkit.

Comparison to current literature (Objective 3) 

In terms of delayed prescribing, previous studies have reported delayed antibiotic prescribing being used by GPs as a compromise to meet patient expectations, maintain relationships between the practitioner and patient, manage uncertainty and provide a safety-net to safeguard against complication, which echoes with our findings (Peters et al., 2011, Ryves et al., 2016). A recent Cochrane Review looked at delayed prescribing and this highlighted that there was no evidence that delayed antibiotic prescribing is safer or more harmful than a no-prescription strategy (Spurling et al., 2013). 

Rowbotham et al (2012) found that NPs are comfortable with a no-prescribing strategy for RTI and have a repertoire of strategies to draw upon when dealing with patients. Previous research has also found that NPs tend to be less likely to yield to patient pressure or conflict, and can acknowledge a patient’s illness without providing prescription (Peters et al., 2011). However, all previous research focussed on delayed antibiotic prescribing in primary care and did not solely focus on the OOH setting.


In this study we met 3 of our 4 objectives.


1: Identified GP and NP experiences of prescribing antibiotics for RTI OOH and explored facilitators and barriers to reducing antibiotic prescribing.

2: Explored similarities and differences between GP and NP antibiotic prescribing and their views on the need for a training intervention and views on content and mode of delivery of such an intervention.

3: Compared OOH findings with existing evidence on in hour antibiotic prescribing. 

The project has just closed and as such we have yet to meet objective 4 (feedback findings to key stakeholders to determine the need for a targeted OOH intervention to optimise prescribing behaviours). We will update SPCR once we have completed Objective 4 and published a peer reviewed manuscript based on the above findings. 

Our findings suggest that existing interventions could be used in OOH care but would need to be tailored to the particular context of OOH and its user population. In particular, organisational factors such as access to patient records, consultation time, working contracts and the amount of audit and/or feedback would be important to consider should a targeted intervention be developed in order to optimise prescribing behaviours in this setting.


Plain English summary

Respiratory tract infections (RTIs) (cough, sore throat, sinusitis, ear infection, and cold) are usually short-lived and are not often serious. Research shows that antibiotics are not very good in treating RTI’s and people usually get better without them. But, antibiotics are often prescribed in primary care out of hours services (6.30pm-8am). This is not best practice because the overuse of antibiotics has been linked to bacterial resistance (which can mean antibiotics are no longer effective when we need them to be). To ensure we still have effective antibiotics in the future unnecessary prescribing needs to be reduced.

Delayed or ‘just in case’ prescribing of antibiotics can help reduce their use. This means the patient is offered a prescription but it is recommended they ‘delay’ taking it for a few days to allow the RTI to improve naturally without medicine/take the antibiotic if the condition worsens. The National Institute for Health and Care Excellence (NICE) recommend that health care professionals who prescribe use this delayed approach. However, research has shown that some GPs and nurses do not like to use this ‘just in case’ approach. In most cases, no antibiotic is needed, but a delayed prescription is likely to provide benefits in comparison to an immediate antibiotic prescription. To date, research has looked at staff who work ‘in-hours’, usually between 7am-7pm. 

Antibiotic prescribing can be influenced by factors that are unique to the out of hours setting. In particular, organisational factors such as access to patient records, consultation time, working contracts and the amount of audit and/or feedback would be important to consider should a targeted intervention be developed in order to optimise prescribing behaviours in this setting.


Published articles

  1. Delayed antibiotic prescribing strategies for respiratory tract infections in primary care: pragmatic, factorial, randomised controlled trialBMJ Little P, Moore M, reader in primary care research, Kelly J, Williamson I, Leydon G, McDermott L, Mullee M, Stuart B, On behalf of the PIPS Investigator

  • Presented at the following conferences: Society of Academic Primary Care South West March 2016, University of Southampton Faculty of Medicine Research Conference June 2016, Society of Academic Primary Care ASM July 2016, GRIN Conference September 2016, NIHR at 10 National Conference November 2016.
  • PPI leaflet in draft for dissemination through PPI channels.
  • Report in draft to be sent to all participants and key stakeholders summarising study findings and requesting feedback on the necessity or otherwise of targeted training for more prudent antibiotic prescribing.

Public involvement

The study team initially had PPI involvement via the local RDS, which was incredibly helpful in developing early lay-summary documentation and focussing the research question. During the project, our PPI rep Dr Karen Postle was an integral member of the research team who attended team and analysis meetings and commented on study paperwork. Overall her input was valuable and pushed us constantly to question how our findings impacted on clinical practice and ‘real life’. As a result of her direct input changes were made to the interview schedule, analysis and write-up.


In the future, it is hoped that the findings from this project will impact on clinical practice through the development of a targeted training in order to optimise prescribing behaviours in the out of hours setting.

Project website: 

This project was funded by the National Institute for Health Research School for Primary Care Research (project number 275)

Department of Health Disclaimer

The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the NIHR School for Primary Care Research, NIHR, NHS or the Department of Health.