Most pharmacists learn clinical development as a sequence of distinct steps. A drug enters Phase 1, moves into Phase 2 if the early evidence is encouraging, advances into Phase 3, and eventually reaches FDA review if the program succeeds.
There is good reason for that structure. Each stage is designed to answer different questions about safety, dosing, efficacy, and benefit versus risk. But the structure also creates something else: time. Trials are designed, sites are opened, patients are recruited, data are collected, results are analyzed, decisions are made, and then much of that infrastructure may need to be assembled again for the next stage.
On September 30, the U.S. Department of Health and Human Services and the Advanced Research Projects Agency for Health, or ARPA-H, launched a new program that asks whether parts of that process can be redesigned much more fundamentally.
The program is called SURPASS, short for Simulation-augmented, Real-time Platform Adaptive Seamless Trials. ARPA-H says clinical development of drugs and biologics often takes more than a decade, costs between $1 billion and $2 billion, and fails about 90% of the time. Its goal for SURPASS is ambitious: develop a clinical-trial system capable of evaluating drugs and biologics in less than four years, at lower cost, with fewer participants and less duplicated infrastructure.
This is not simply another federal effort to make clinical trials a little faster. ARPA-H is questioning whether the traditional stop-and-start structure of development itself can become more continuous.
For pharmacists, that matters because the way evidence is generated around future medications may begin to look very different from the clinical-development model many of us learned in school.
What if a clinical trial did not have to keep starting over?
One of the biggest ideas behind SURPASS is the platform trial.
A conventional drug-development program might run one trial, complete it, analyze the results, design the next trial, negotiate with sites again, activate those sites, build the necessary systems, recruit another population, and begin again.
A platform trial attempts to preserve much more of that infrastructure.
FDA describes platform trials as studies that can evaluate multiple products on an ongoing basis, with treatments entering or leaving the platform over time. Shared elements can include control groups, clinical sites, oversight structures, data systems, and portions of the underlying trial protocol. FDA has noted that well-designed master protocols can generate reliable safety and effectiveness evidence while reducing some of the duplication associated with running completely separate studies.
SURPASS wants to take that idea considerably further.
ARPA-H envisions what it calls perpetual, adaptive platform trials. Instead of repeatedly dismantling a trial infrastructure after answering one question, a platform could potentially remain active while new treatment arms are introduced, ineffective strategies are removed, and evidence continues accumulating within a common system.
That does not mean a clinical trial would simply change direction whenever investigators felt like it. Adaptive trials still require rigorous statistical planning. FDA defines adaptive designs around prospectively planned modifications based on accumulating trial data, and its guidance emphasizes that those designs must still produce reliable and interpretable evidence.
The difference is that researchers may not always have to wait until the entire study is over before learning from what is happening inside it.
SURPASS is trying to make the trial learn while it is running
ARPA-H divides the program into three major technical areas, and together they show how different this future could look.
The first is what the agency calls a “phaseless design engine.” The goal is to integrate predictive computational models, including digital twins and other simulations, into trial design so researchers can model clinical and operational outcomes before a trial launches. ARPA-H also wants researchers to build enough evidence around these approaches that regulators can develop confidence in when and how they should be used.
The word “phaseless” needs some context. SURPASS is not an announcement that Phase 1, Phase 2, and Phase 3 requirements have disappeared. It is a research program exploring whether some of the artificial stops between development stages can be reduced through more continuous, seamless trial designs while still generating the evidence required for regulatory decision-making.
Seamless trials already exist in certain forms. They can combine stages of development within one protocol so that, under predefined rules, a study transitions from one objective to the next without shutting down enrollment and rebuilding the trial from scratch. The concept is not to eliminate scientific questions, but to answer them within a more continuous structure.
SURPASS’s second technical area is a continuous inference engine. ARPA-H wants statistical methods that can analyze trial data in real time or on demand while maintaining valid conclusions. If successful, those systems could allow researchers to make certain predefined adaptations earlier, identify ineffective strategies sooner, and potentially reduce the size of conventional control groups by making greater use of shared controls and predictive modeling.
The third area may be less scientifically glamorous, but it could be just as important: operations.
Clinical trials do not take years only because biology is difficult. They also take years because activating sites, negotiating agreements, cleaning data, creating datasets, onboarding new treatment arms, coordinating participants, and managing dozens of operational processes take enormous amounts of time.
ARPA-H wants to develop an “agentic operations layer” that uses automation to reduce some of that administrative burden. The program specifically identifies trial startup, data collection, data cleaning, dataset construction, and the addition of new treatment arms as areas where automation could make the system move faster.
Put those pieces together and SURPASS begins to look less like a traditional clinical trial and more like a continuously operating evidence-generation system.
This is bigger than another story about AI in drug development
Pharmacy Unlocked has followed several recent developments involving AI moving deeper into healthcare. We have looked at AI entering the patient chart, pharmaceutical R&D, scientific modeling, and the generation of preclinical evidence.
SURPASS adds another layer.
The question is no longer only whether AI can help scientists discover a molecule or analyze a dataset. It is whether computational modeling, real-time statistics, automation, shared infrastructure, and adaptive design can change the architecture of clinical development itself.
That distinction matters.
A clinical trial is the mechanism through which we learn whether a therapy actually works in people and whether its benefits justify its risks. If the structure used to generate that evidence becomes more adaptive, more computational, and more continuous, nearly every function around drug development will have to evolve with it.
Biostatisticians will need to build and validate more sophisticated analytical methods. Clinical operations teams will need to manage trials where treatment arms can enter and exit a platform. Regulatory professionals will need to understand how evidence generated through novel designs supports approval decisions. Data teams will need to integrate information quickly enough for near-real-time analysis. Investigators will need to understand predefined adaptation rules. Technology teams will need to build systems capable of supporting all of it without compromising data integrity.
Pharmacists working in industry will increasingly operate somewhere inside that environment.
The trial infrastructure around the patient is changing too
SURPASS is only one part of what HHS announced on September 30.
ARPA-H also launched three complementary projects aimed at problems that sit around the trial itself. STACK is intended to expand U.S. clinical-trial capacity by using AI to accelerate site activation and help research-naïve healthcare locations become capable clinical-trial sites. COMMONS is designed to create a privacy-focused data infrastructure that could support access to regulatory-grade data at national scale. CINCH focuses on helping patients contribute real-world data, improve care navigation, and connect more quickly with appropriate clinical trials.
That broader effort is important because clinical-trial efficiency is not simply a statistical problem.
A beautifully designed adaptive trial still cannot run if sites take months to activate. A predictive model cannot help much if the necessary data are trapped in disconnected systems. A promising study cannot recruit if eligible patients cannot find it or participation is too burdensome.
HHS is effectively approaching the clinical-trial ecosystem as a connected system: trial design, data, operations, clinical sites, and the patient experience all have to improve together.
For pharmacists, that widens the career implications even further.
Clinical development careers may become much more interdisciplinary
A pharmacist interested in pharmaceutical industry often hears about a familiar group of roles: medical affairs, regulatory affairs, clinical development, pharmacovigilance, medical information, and clinical operations.
Those remain important career areas. But what happens inside them is changing.
A clinical-development pharmacist may increasingly need to understand why an adaptive design was selected, how a shared control group works, or how a model influences dose or enrollment decisions.
A regulatory professional may need to evaluate whether a novel statistical approach generates evidence robust enough for an FDA submission.
A medical affairs professional may eventually need to explain evidence produced through a platform trial that looks very different from the classic parallel-group randomized trial healthcare professionals are accustomed to reading.
A drug-safety professional may work in an environment where safety information is being evaluated continuously while the trial itself adapts.
A clinical-operations professional may help manage a platform in which new investigational therapies can be added without rebuilding the entire study infrastructure.
A pharmacist working in research strategy may need to understand where computational modeling can improve a development plan and where traditional human evidence remains indispensable.
That does not mean every PharmD needs advanced training in Bayesian statistics or machine learning.
It does mean that understanding how modern evidence is created may become increasingly valuable across many pharmaceutical careers.
The pharmacist who knows how to read the final clinical trial paper will always have value. The pharmacist who also understands how the trial was designed, why it adapted, what the model contributed, what assumptions were made, and whether the resulting evidence is trustworthy may bring something even more useful.
Faster cannot mean less rigorous
There is an important caution inside all of this.
A four-year clinical-development target sounds exciting, but speed cannot become the primary standard for evidence.
SURPASS itself recognizes that problem. ARPA-H repeatedly emphasizes that the goal is to accelerate decisions while maintaining rigorous evidence generation and building regulatory confidence in the new approaches.
That will be one of the hardest parts.
The traditional clinical-development system may be slow, but many of its safeguards exist for a reason. Randomization matters. Appropriate controls matter. Prespecified endpoints matter. Statistical validity matters. Safety follow-up matters. Regulatory review matters.
Adaptive designs introduce additional complexity because decisions made during a trial can potentially influence the statistical properties of the evidence. That is why FDA guidance requires adaptations to be carefully planned and evaluated so the final results remain interpretable.
Computational models create similar questions.
A digital twin may be able to simulate potential patient outcomes, but how accurate is it? Which patients were represented in the data used to build it? Under which conditions does it fail? Can its predictions be independently validated? When can simulated information supplement a control group, and when is a conventional human comparator still necessary?
Those questions are not reasons to avoid innovation.
They are the work required to make innovation trustworthy.
And that may be one of the most important career themes emerging across everything Pharmacy Unlocked has covered recently: as technology becomes better at generating information, the ability to evaluate the quality of that information becomes more valuable.
Do not just “learn AI.” Learn one changing workflow.
For pharmacists interested in industry, the practical takeaway from SURPASS is not to spend the weekend trying to become an expert in clinical-trial statistics.
Choose one concept and understand it well enough to explain why it exists.
Start with an adaptive trial. Learn how predefined rules can allow a study to change based on accumulating evidence.
Or study platform trials and understand why multiple therapies might share one infrastructure and control group.
Learn what a seamless trial is and why eliminating the downtime between traditional phases could shorten development.
Look into digital twins and predictive modeling and ask what evidence would be required before simulated patients could meaningfully influence a clinical decision.
Explore real-world data and why connecting information generated outside traditional trials could affect recruitment, modeling, or follow-up.
Then take the career step.
Search the term on LinkedIn Jobs or pharmaceutical company career pages. Do not search only for “pharmacist.” Search “adaptive clinical trial,” “platform trial,” “clinical development,” “biostatistics,” “trial innovation,” “regulatory science,” or “clinical operations.”
Open the roles that appear and look for patterns. Which companies are hiring? Which departments own this work? What backgrounds do they accept? Which technical skills appear repeatedly? Where does a PharmD overlap naturally, and where would additional training be required?
That is a much better way to learn an emerging field than choosing another credential based on a generic list of certifications for pharmacists.
You are studying the work first.
The clinical trial may become a living system
SURPASS is still a research program. Its less-than-four-year target is a goal, not a result that has already been achieved. The program currently has an open funding opportunity, and ARPA-H is seeking multidisciplinary teams across statistics, AI, clinical-trial design, operations, and regulation to build and test the proposed system.
There will be technical challenges. There will be regulatory challenges. Some ideas may work better in certain diseases than others. Conventional trials will not suddenly disappear.
But the direction is worth understanding.
The clinical-development model pharmacists learned as a clean progression from Phase 1 to Phase 2 to Phase 3 may increasingly be supplemented by trials that are more continuous, more adaptive, more computational, and built on infrastructure that can persist across multiple therapies.
That could change how quickly ineffective drugs are abandoned. It could change how control groups are constructed. It could change how patients are recruited. It could change how evidence is analyzed. It could change how pharmaceutical teams are organized around clinical development.
Most importantly, it could change the skills required of the people working around future medicines.
Pharmacists do not need to become the statisticians designing every adaptive algorithm or the engineers building every digital twin. But they should understand the system those tools are entering.
Because clinical trials are not simply a step drugs pass through before approval.
They are where the evidence underlying medication decisions is created.
If the architecture for creating that evidence changes, pharmacists should understand the change early.
That is not simply following where drug development is going.
It is learning where the next generation of pharmaceutical careers may be going with it.
Resources & Citations
- U.S. Department of Health and Human Services. “HHS Launches SURPASS and New Efforts to Accelerate Faster, Smarter Clinical Trials.” September 30, 2026. Primary federal announcement describing SURPASS, the more-than-decade clinical-development timeline, development costs, failure rates, the less-than-four-year ambition, predictive computational models, shared infrastructure, common controls, real-time analysis, and the complementary STACK, COMMONS, and CINCH projects.
- Advanced Research Projects Agency for Health. “SURPASS: Simulation-augmented, Real-time Platform Adaptive Seamless Trials.” September 30, 2026. Official program page describing the current clinical-development problem, the program’s goal of shortening trials to under four years, perpetual adaptive platform trials, three technical areas, and the open funding opportunity.
- U.S. Food and Drug Administration. “Adaptive Design Clinical Trials for Drugs and Biologics: Guidance for Industry.” December 2019. FDA guidance describing the principles for appropriately designing, conducting, analyzing, and reporting adaptive clinical trials intended to support drug and biologic development.
- U.S. Food and Drug Administration. “E20 Adaptive Designs for Clinical Trials.” Draft Guidance. September 2025. FDA and ICH guidance addressing the planning, conduct, analysis, and interpretation of adaptive clinical trials intended to provide reliable evidence supporting efficacy and benefit-risk assessment.
- U.S. Food and Drug Administration. “Guidance Recap Podcast: Master Protocols for Drug and Biological Product Development.” 2026. FDA explanation of platform, umbrella, and basket trials and the potential efficiencies created by shared controls, clinical-site infrastructure, governance, and other common trial elements.
- U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation. “Transformation of the Clinical Trial Enterprise: Lessons Learned from the COVID-19 Pandemic.” 2024. Federal report describing adaptive trials, platform trials, master protocols, shared infrastructure, and the potential benefits and operational complexities of newer trial designs.
- SAM.gov. “Simulation-augmented, Real-time Platform Adaptive Seamless Trials (SURPASS), ARPA-H-SOL-26-164.” September 30, 2026. Official federal contracting notice confirming the active SURPASS research opportunity and ARPA-H’s solicitation of teams to develop the program’s technical and regulatory framework.
