General Practitioners' antibiotic prescribing decisions for children with upper respiratory tract infection: a Think-Aloud study

Talk Code: 

The problem

Over-prescribing of antibiotics is widespread and poses significant public health problems. Upper respiratory tract infection (URTI) has been identified as a target area for reducing antibiotic use in general practice. Interventions aimed at improving prescribing practices may be more effective when based on a detailed understanding of processes underlying General Practitioners' (GPs') prescribing decisions. This study used the ‘Think-Aloud' method, where participants are asked to complete a task whilst verbalising their thoughts, to investigate how GPs make prescribing decisions.

The approach

In individual, face-to-face interviews, GPs responded to seven patient scenarios by thinking aloud while making prescribing decisions and rating how difficult they found decision-making to be. The scenarios were originally developed for an online, Scotland-wide study of prescribing decision-making, and described children with acute sore throat or acute otitis media. The scenarios were constructed to include features representative of real patients, such as relevant signs and symptoms, and non-clinical factors such as parental preferences. GPs who had participated in the online study were recruited via email. Interviews were audio-recorded and transcribed verbatim. Interview transcripts were coded for the information used in the decision process and/or related to decision difficulty. An inductive approach identified further emergent themes.


Five GPs based across Central, East and West Scotland were recruited (3 male and 2 female, with 2 to 24 years' experience as a GP). Information used to make decisions included illness severity, duration of illness, and parental preferences and concerns. Decisions were more difficult when scenarios included the following factors: longer duration of illness, follow-up consultations, and parents who were worried, had attempted symptomatic relief, wanted antibiotics, and mentioned the impact of illness on their lives. Longer duration of illness was a key influence on decision-making, resulting in variation in judgments and decisions. Emergent themes included variation across GPs in: a) interpretation of clinical features and corresponding guideline recommendations, and b) emphasis on the importance of shared decision-making. Delayed prescriptions were often selected to represent a shared decision with the parent, or a compromise between clinical signs and symptoms and parental comments.


Antibiotic prescribing decisions depend on interpretations of patient information and associated guideline recommendations, which may vary across GPs. Interventions aimed at improving prescribing practices may be more effective when tailored to take this variation into account, or when focussed on situations where decision difficulty is greater. Delayed prescriptions are already advocated as a method of reducing inappropriate prescribing: it may be beneficial to highlight, to the wider GP population, the usefulness of this method for facilitating shared decision-making. Due to the relatively small sample in this study, further research into antibiotic prescribing decision processes would be beneficial.


  • Nicola McCleary, City University London, London, England, UK
  • Marion Campbell, City University London, London, England, UK
  • Craig Ramsay, City University London, London, England, UK
  • Jill Francis
  • Julia Allan, City University London, London, England, UK