Randomized Controlled Trial Design Based on Patient Cohorts： Methods and Applications of Trials Within Cohorts / 中国实验方剂学杂志

ABSTRACT

Trials within cohorts （TwiCs） are design methods derived from randomized controlled trials （RCTS）. They have been widely used in chronic disease areas such as tumors and cardiovascular diseases. The basis of the TwiCs design is a prospective cohort of specific diseases. When RCTS need to be implemented， some patients meeting the inclusion and exclusion criteria are randomly sampled from the cohort to receive "trial interventions"， while the remaining patients in the cohort who meet the inclusion and exclusion criteria continue to receive conventional treatment as control groups. By comparing the efficacy differences between the intervention measures of the trial group and the control group， the efficacy of intervention measures was evaluated. Within the cohort， the same process could be repeated to carry out multiple RCTS， so as to evaluate different intervention measures or compare the efficacy of different doses or timing of interventions. Compared with classical RCTS， TwiCs make it easier to recruit patients from the cohort and have higher external validity， providing a new research paradigm for improving the efficiency and applicability of RCTS in clinical practice. However， TwiCs may also face the challenge of poor compliance of patients in the cohort. Researchers need to take effective measures to control these patients in the design and operation of TwiCs. This article focused on the methodological key points during the implementation of TwiCs， including multi-stage informed consent （patients are informed of consent at three stages： entering the cohort， entering the trial group， and after the trial）， randomization procedures （only random sampling of patients from the cohort to receive "trial interventions"）， sample size calculation， and statistical analysis methods. The article also compared the differences between TwiCs and traditional RCTS and illustrated TwiCs research design and analysis with examples， so as to provide new research ideas and methods for clinical researchers.