Target Trial Emulation brings the logic of randomized trials into real-world data

Can registry data alone tell us which treatments are effective? Is it possible to know whether a drug prevents heart attacks—or increases the risk of serious side effects—without conducting a randomized trial? How can Target Trial Emulation be used?

Target Trial Emulation (TTE) is a research method that allows researchers to explore causal relationships from observational data, such as patient registries or electronic health records, provided that key assumptions – such as no unmeasured confounding – are met. TTE does not replace randomized controlled trials (RCTs), but it applies their logic and rigor to real-world data analysis.

Senior Data Scientist Johanna Vikkula from Medaffcon explored the topic at ISPOR last autumn.

“TTE begins with the question: if this were a clinical trial, how would I design the study? From there, you can start outlining the research steps. How do I identify the right patients? How do I define their follow-up time? Are there certain criteria that should exclude a patient from the study?” Vikkula explains.

The researcher builds an observational study

The foundation of TTE is that the researcher designs the observational study as if it were a randomized controlled trial. The structure of the analysis is based on a hypothetical “target trial.” This includes defining who is studied, what the intervention is, what the comparator is, when exposure begins, how long the follow-up lasts, and what the primary outcome is. It also states if randomization is needed.

This framework is then emulated using existing data sources, such as national registries or electronic health records. The goal is to replicate the structure of an RCT as closely as possible and minimize bias. Emulation means reconstructing the design and logic of a randomized trial using observational data.

Traditional observational studies may be prone to several biases. For example: Confounding occurs when patients who receive treatment differ in systematic ways that also affect outcomes. Immortal time bias refers to errors in timing that can lead to misleading conclusions about treatment effects.

TTE helps distinguish correlation from causation, simulate treatment effects, and generate effectiveness data when RCTs are not available.

Barbara Dickerman’s statin study is one of the best-known TTE examples

One of the best-known and most cited examples of Target Trial Emulation is the study led by Barbara Dickerman.

“The study looked at the link between statins and cancer risk, and it clearly shows how observational analysis can go wrong without the structure of a randomized trial,” Vikkula says.

Earlier observational studies had suggested a protective effect of statins, while clinical trials did not find similar association. When applying TTE in the observational studies, researchers showed that these findings were due to immortal time bias and other analytical errors. When the analysis was reconstructed as though it were an RCT, the protective effect disappeared.

Target Trial Emulation can also be used in pharmaceutical research

TTE is also applicable in industry-sponsored studies. It can be used to:

• Provide additional evidence of a drug’s real-world effectiveness
• Prepare for health technology assessments (HTA)
• Support registry-based decision-making when conducting a new RCT is not feasible