Computable Phenotype Implementation for a National, Multicenter Pragmatic Clinical Trial: Lessons Learned From...

Area of Science:

• Cardiovascular Clinical Trials
• Health Informatics
• Clinical Trial Recruitment

Background:

• Large-scale cardiovascular trials face high costs and low patient enrollment.
• Computable phenotypes, using algorithms on electronic health records, are crucial for efficient recruitment.
• This paper details the computable phenotype development for the ADAPTABLE trial.

Purpose of the Study:

• To describe the development and implementation of a computable phenotype for the ADAPTABLE trial.
• To assess the efficiency of computable phenotypes in recruiting for pragmatic clinical trials.
• To identify key lessons learned during the computable phenotype development and implementation process.

Main Methods:

• A computable phenotype was developed to identify adults meeting ADAPTABLE trial eligibility criteria.
• The phenotype identified over 650,000 potentially eligible patients across 40 sites.
• 15,076 participants were enrolled between April 2016 and June 2019.

Main Results:

• The computable phenotype successfully identified a large cohort of eligible patients.
• Patient enrollment in the ADAPTABLE trial reached 15,076 participants.
• Variability in source data quality and local coding patterns necessitated site-specific validation and modification.

Conclusions:

• The ADAPTABLE computable phenotype proved an efficient recruitment tool for a multisite pragmatic trial.
• The development and implementation process offers valuable insights for future large-scale trials.
• Sustained multidisciplinary collaboration is essential for successful computable phenotype implementation.