Cluster randomized trials utilizing primary care electronic health records: methodological issues in design, conduct,...

Area of Science:

• Health Services Research
• Clinical Trials Methodology
• Primary Care Research

Background:

• Growing interest in utilizing electronic health records (EHRs) for clinical and cluster randomized trials.
• This study details methodological challenges encountered in two cluster randomized trials implemented within the Clinical Practice Research Datalink (CPRD) EHRs.
• Trial registrations: ISRCTN 47558792 and ISRCTN 35701810.

Purpose of the Study:

• To report on methodological issues during the implementation of two cluster randomized trials using EHRs.
• To evaluate the feasibility and challenges of conducting trials in primary care settings via EHRs.
• To provide insights for future EHR-based trial design and execution.

Main Methods:

• Two cluster randomized trials conducted in primary care: one on antibiotic prescribing for respiratory infections, another on stroke secondary prevention adherence.
• Data derived from documentary records and trial datasets, focusing on research ethics, governance, recruitment, allocation, sample size, intervention implementation, and analysis.
• Utilized electronic health records from the Clinical Practice Research Datalink (CPRD).

Main Results:

• Research governance approvals obtained from over 150 primary care organizations across England, Wales, and Scotland.
• Recruited 104 general practices for the antibiotic trial and 106 for the stroke trial within six months.
• Interventions were remotely installed; participant numbers varied significantly (mean 5,588 for antibiotics, 110 for stroke).
• Outcome measures showed strong correlations (e.g., 0.91 for antibiotic prescribing), highlighting the need for careful eligibility definition.

Conclusions:

• Cluster randomized trials can be efficiently conducted with large samples from UK general practices using EHRs.
• Geographical dispersal of trial sites poses challenges for research governance and intervention implementation.
• Pretrial data analysis is crucial for informing robust trial design and analysis plans.