Emulated target trials: Trial-grade evidence from real-world observational data

What are emulated target trials and why do they matter?

Randomised controlled trials (RCTs) remain the gold standard for establishing causal evidence in biomedical research, but they are costly, time-consuming, and often limited in scope. Once a drug or technology is approved and used in the real world, new questions arise: does it work equally well in different patient populations? What are the long-term outcomes? How does it compare against alternative treatments already on the market? What are the socio-economic costs and benefits to patients and healthcare funders?

This is where emulated target trials (ETTs) come in. First proposed nearly a decade ago by researchers at the Harvard School of Public Health, ETTs provide a structured framework for deriving causal insights from real-world data (RWD). The idea is simple but powerful: design the protocol for the trial you wish you could run, and then emulate it using observational data such as electronic health records, claims databases, or disease registries. By aligning inclusion and exclusion criteria, treatment allocation, time zero, and outcomes to mirror an RCT, ETTs produce evidence that is more rigorous than standard observational studies. In doing so, they hold significant promise for accelerating precision medicine, improving policy decisions, and generating evidence that is both timely and relevant.

‍

Why are they not commonplace?

Despite their potential, ETTs are still underutilised in pharmaceutical and biomedical research, for a number of reasons:

• Regulatory uncertainty. While regulators are increasingly open to real-world evidence, there is no harmonised guidance on how ETTs should be evaluated in submissions.
• Data limitations. High-quality, longitudinal RWD is rare. Missing variables, inconsistent coding, and unlinked datasets make trial emulation difficult.
• Skills gap. Designing an ETT requires expertise in handling time alignment, immortal time bias, dynamic treatment regimes and time-varying covariates; skills not yet widespread in industry.
• Complexity and cost. Compared with simpler observational analyses, writing a full trial protocol and implementing it with RWD is more resource intensive.
• Cultural conservatism. RCTs remain the familiar and lower-risk route for pivotal evidence. ETTs challenge ingrained hierarchies around how evidence should be generated.
• Lack of standards. Unlike RCTs (CONSORT) or observational studies (STROBE), ETTs lacked universally accepted reporting guidelines until the very recent publication of the TARGET statement, making results harder to compare and trust.

‍

When and where are ETTs used?

Although not mainstream, ETTs are being adopted in specific contexts where RCTs are impractical, unethical, or too costly:

• External control arms. In rare diseases and oncology, ETTs provide comparators for single-arm trials, saving time and improving patient acceptability.
• Label expansion. They help test existing drugs in new indications, generating evidence to support or prioritise future RCTs.
• Post-marketing evaluation. ETTs are used to monitor long-term safety and effectiveness, capturing outcomes beyond the scope of pivotal trials.
• Health economics and outcomes research (HEOR). By estimating comparative effectiveness in broader populations, ETTs strengthen value dossiers for payers and Health Technology Assessment agencies.
• Geographic bridging. They help assess whether trial results from one region apply in other regions, supporting local reimbursement decisions.
• Precision medicine. By exploring treatment heterogeneity, ETTs can identify which subgroups benefit most, a step towards more personalised care.

In these use cases, regulators have shown willingness to consider ETT evidence, particularly when RCTs are infeasible.

‍

Key considerations when planning and conducting an ETT

Running an ETT is not as simple as “plugging data into a model.” To ensure credibility, several principles must be respected:

1. Write the trial protocol first. Define eligibility, interventions, outcomes, and analysis as if running a real trial. Only then test whether your data can support it.
2. Mind time zero. Eligibility criteria and treatment allocation must align at baseline to avoid immortal time bias.
3. Beware data gaps. Missing or inaccurate data can invalidate the emulation. Not all RWD is fit for purpose.
4. Test assumptions. Use methods like negative control outcomes to check for evidence of residual confounding.
5. Benchmark against RCTs. Consistency with known trial results builds confidence in ETT findings.
6. Plan for adoption. Even the best-designed ETT will fail to influence policy if stakeholders are not trained to interpret and use the evidence.

Like RCTs, ETTs involve trade-offs: they are narrower in scope but deliver stronger causal inference than most observational designs. Success depends on careful design, high-quality data, and transparent reporting.

‍

In conclusion: Emulated target trials are not a silver bullet, but they represent one of the most rigorous frameworks we have for unlocking the value of real-world observational data. By bridging the gap between RCTs and messy clinical reality, they offer a path to more timely, relevant, and cost-effective evidence generation. For organisations willing to invest in the necessary skills and infrastructure, ETTs can provide not only scientific insight but also a strategic advantage.

‍