Should we stop conducting prospective randomized trials now that target trial emulation is available?

Target Trial Emulation (TTE) has emerged as a powerful methodology for estimating causal effects from observational data, offering a cost-effective and time-efficient alternative to randomized controlled trials (RCTs). By structuring real-world data (RWD) to mirror the design of an RCT, TTE allows for rigorous comparative effectiveness research without the need for prospective data collection. However, despite its growing application, TTE has inherent limitations that, in many cases, necessitate traditional prospective trials.This paper examines the circumstances in which prospective RCTs remain indispensable, focusing on six key factors: (1) Data Availability and Quality, where insufficient RWD precludes meaningful trial emulation; (2) Electronic Health Records (EHR) Limitations, which introduce biases due to incomplete or inconsistent data capture; (3) Unmeasured Confounding, where key clinical variables are missing from observational datasets, limiting causal inference; (4) Outcome Measurement Bias, particularly in studies requiring standardized endpoints; (5) Time-Varying Exposures and Immortal Time Bias, where temporality issues can distort causal estimates; and (6) Generalizability, as observational datasets may lack diversity, reducing external validity.Using clinical examples from cardiology, oncology, mental health, and perioperative medicine, this paper underscores that while TTE can enhance research efficiency, it cannot fully replace prospective trials. We propose a hybrid framework where TTE informs trial design but is supplemented by RCTs in high-stakes settings requiring precise effect estimation. Ensuring rigorous study feasibility assessments before choosing between TTE and RCTs is crucial for maintaining scientific integrity and guiding clinical decision-making.