The Endpoint Bubble: Why Clinical Trials Are Drowning in Measurements That Don’t Matter
By Vlad Bogin, MD, FACP CEO, Cromos Pharma
One of the more noticeable shifts in clinical development over the past decade has been the steady expansion of endpoints in clinical trial protocols. What used to be a focused set of primary and secondary measures has grown into long lists of biomarkers, imaging readouts, digital signals, patient‑reported outcomes, pharmacodynamic markers, and exploratory mechanistic assessments. In some protocols, the endpoint section now runs several pages.
The motivation is understandable. Patient recruitment is difficult, development costs are high, and sponsors want to extract as much information as possible from every participant. Advances in diagnostics and digital health have also made it easier to measure aspects of disease that were previously inaccessible.
But more measurement does not automatically translate into more insight. At a certain point, the accumulation of endpoints stops clarifying a study’s purpose and begins to obscure it.
Why Sponsors Keep Adding Endpoints
When used thoughtfully, additional endpoints can be valuable. They help characterize the full therapeutic profile of an investigational product, reveal mechanistic signals, support future development decisions, and generate hypotheses for subsequent studies.
In oncology, overall survival may remain the definitive endpoint, but progression‑free survival, objective response rate, duration of response, minimal residual disease, and quality‑of‑life measures all contribute to understanding treatment effect. In cardiovascular and kidney disease, biomarkers and imaging often help explain why a therapy succeeds or fails.
A Note on Composite Endpoints
Composite endpoints deserve special attention. When well‑constructed, they can increase event rates, reduce sample size requirements, and capture the full clinical relevance of a therapy. But poorly designed composites can obscure interpretation, especially when component events differ in clinical importance or occur at very different frequencies. The goal is not to avoid composites, but to ensure they reflect a coherent clinical construct and support the study’s primary decision.
The issue is not the existence of these endpoints. It is the lack of discipline in selecting which ones truly matter.
The Hidden Costs of Endpoint Expansion
Recent analyses show that the number of endpoints in Phase III trials has risen sharply, increasing by 37% between 2015 and 2021, according to Tufts CSDD data published in Applied Clinical Trials. This growth reflects both scientific opportunity and a growing tendency to measure more simply because we can.
- Statistical Complexity
Regulators have been clear: testing multiple endpoints increases the risk of false‑positive findings unless multiplicity is properly controlled. The FDA’s 2024 guidance on multiple endpoints emphasizes the need for prespecified hierarchies and transparent inferential strategies. Without this structure, a trial can generate more confusion than clarity.
- Operational Burden
Every additional endpoint adds work. More procedures, more site training, more monitoring, more data queries, and more opportunities for deviations and missing data. Analyses from the Tufts Center for the Study of Drug Development show that roughly a quarter of procedures in modern protocols do not support primary or key secondary endpoints. In our own protocol reviews, we often see similar patterns: assessments that are unlikely to influence decisions but still add friction to study conduct.
- Patient Burden
Longer visits, more questionnaires, more imaging, and more blood draws all affect the patient experience. In chronic disease and oncology trials, this burden can directly influence retention.
- Interpretability
Perhaps the most important consequence is the one least discussed. Trials with dozens of endpoints often produce a patchwork of mixed results. Some endpoints are positive, others neutral, others contradictory. Instead of a clear answer to the original scientific question, teams are left debating which signals matter most.
A clinical trial should tell a coherent story. Too many endpoints make that story harder to interpret.
A More Useful Way to Think About Endpoints
At Cromos, we review a large number of protocols each year. Across these reviews, a consistent pattern emerges: many exploratory endpoints are included “just in case,” without a clear link to a development decision. When we ask teams whether removing a particular exploratory endpoint would change dose selection, go/no‑go criteria, indication prioritization, or regulatory strategy, the answer is often no.
This is not a criticism of exploratory science. Exploratory endpoints are essential for understanding biology and informing future studies. But they should be chosen deliberately, not simply because the technology exists to measure them.
Endpoint Discipline: A Practical Approach
The most effective protocols start with a simple question: What decision is this study intended to support?
From there:
- The primary endpoint should capture the single most important measure of success. If it is positive, the study has achieved its fundamental objective.
- Key secondary endpoints should address questions that materially influence clinical interpretation, regulatory decision‑making, reimbursement, or adoption into practice.
- Exploratory endpoints should remain exploratory. They should generate hypotheses, not carry the weight of confirmatory evidence. And they should be included only if they have a plausible path to influencing future development.
This approach aligns closely with the principles outlined in ICH E9(R1), which emphasizes designing trials around a clearly defined scientific question and estimand rather than maximizing the number of measurements collected.
A Simple Test We Use When Reviewing Protocols
When our team evaluates a protocol, we apply three straightforward questions:
- Can the primary objective be summarized in one sentence?
- Does each secondary endpoint answer a distinct, clinically meaningful question?
- Would removing an exploratory endpoint change a development decision?
If the answer to the last question is no, that endpoint probably does not belong in the protocol.
This is not about limiting scientific curiosity. It is about ensuring that every measurement serves a purpose.
Looking Ahead
As data collection technologies continue to advance, the temptation to measure everything will only grow. But the value of a clinical trial lies not in the volume of data collected, but in the clarity of the conclusions it enables.
Endpoint discipline is not about doing less. It is about doing what matters. Teams that embrace this mindset will run cleaner, more interpretable studies and make better decisions as a result.
References
- FDA. Multiple Endpoints in Clinical Trials: Guidance for Industry. 2024.
- Unger JM, Mazza GL, Elsaid MI, et al. When to adjust for multiplicity in cancer clinical trials. JNCI Monogr. 2025;68:3‑9.
- Kahan BC, Hindley J, Edwards M, Cro S, Morris TP. The estimands framework: a primer on the ICH E9(R1) addendum. BMJ. 2024;384:e076316.
- Baracaldo‑Santamaría D, Feliciano‑Alfonso JE, Ramirez‑Grueso R, et al. Making sense of composite endpoints in clinical research. J Clin Med. 2023;12:4371.
- Getz KA, Campo RA. Variability in protocol design complexity by phase and therapeutic area. Ther Innov Regul Sci. 2017;51:593‑603.
- Getz KA. Protocol design trends and their impact on clinical trial performance. Tufts CSDD Impact Report. 2022.
- Favaro MA, Sassano C, Sherafat R. Enabling evidence‑based study endpoint selection. Applied Clinical Trials. April 19, 2024.
