Article Text

PP36 Use of Electronic Health Records to Implement a Cluster Randomised Trial in Primary Care
  1. M Ashworth2,
  2. J Charlton1,
  3. P Little3,
  4. M Moore2,
  5. L Yardley4,
  6. M Gulliford4,
  7. A Dregan5,
  8. T van Staa6,
  9. L McDermott7,
  10. G McCann1
  1. 1National Institute for Health Research (NIHR) Biomedical Research Centre (BRC) at Guy’s and St Thomas’ Hospital, King’s College, London, UK
  2. 2Clinical Practice Research Datalink, Modern Humanities Research Association (MHRA), London, UK
  3. 3Department of Psychology, University of Southampton, Southampton, UK
  4. 4Department of Primary Care and Public Health, King’s College, London, UK
  5. 5Department of Primary Care, University of Southampton, Southampton, UK
  6. 6Department of Primary Medical Care, University of Southampton, Southampton, UK
  7. 7Department of Psychology, University of Southampton, Southampton, UK


Background Electronic health records (EHRs) represent a potentially valuable resource for evaluating health interventions. This research aimed to evaluate the feasibility of implementing a cluster randomised trial in a primary care database. The trial intervention aimed to reduce antibiotic prescribing in respiratory tract infections (RTIs).

Methods The trial was implemented using the Clinical Practice Research Datalink (CPRD). All CPRD general practices in England and Scotland were invited to participate. Allocation was stratified by region/country and practice list size. Non-trial practices provided an external comparison group. Interventions were electronic reminders, activated during consultations for respiratory infections by persons aged 18 to 59 years, installed remotely at intervention practices. Outcomes at 12 months were evaluated through analysis of EHR data collected into CPRD with linked HES. The primary outcome was the proportion of consultations for RTI with antibiotics prescribed.

Results There were 104 (53 Intervention, 51 Control) CPRD general practices were recruited to the study. Individual participants were Intervention, 294,259; Control, 264,065; and External Comparison, 518,315. Interventions were successfully installed at all intervention practices. Banners offering accessing to the intervention material were regularly viewed, but prompts were actively viewed at only 47 per 1,000 consultations for sore throat. The median (IQR) practice-specific proportion of RTI consultations with antibiotic prescribed were: Intervention, 53 (45, 57) Control, 54 (45, 59); External comparison 54 (45, 62). The linkage of CPRD with HES suggested no increase in hospital admissions following the intervention.

Conclusion Cluster randomised trials may be implemented successfully, with large numbers of participants, using a primary care electronic database. Remote delivery of interventions to practices is feasible but future development of more effective interventions is required.

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.