Written by Johnathon Anderson, Ph.D., a research scientist, Associate Professor & Program Officer at the University of California Davis School of Medicine, and CEO of Peptide Systems
Published by: Peptide Systems
Executive Summary
The Shift: In 2024, China officially surpassed the US in annual clinical trial starts (7,100 vs. 6,000), driven by regulatory reforms that function as "implied approval."
The Strategy: China is executing a game-theoretical strategy of "Serendipity Maximization." By lowering the "Cost per Experiment" by ~40–60%, they are industrializing phenotypic screening at a population scale.
The Data: A 2023 analysis by Arash Sadri in J. Med. Chem. confirms that only 9.4% of approved drugs originate from "Target-Based Design," exposing the inefficiency of the Western "Rational Design" dogma.
The Solution: To compete, the US must adopt "Clinical Abundance" policies (Institute for Progress) and leverage the "Biotech Social Contract" (Kolchinsky) to lower the transaction costs of innovation.
In recent years, one-third of US Biopharma acquisitions have come from China. And there is another haunting statistical coincidence at the heart of modern drug development.
In clinical trials, researchers define statistical significance as a P-value of 0.05 (a 5% probability that the result is random noise). According to a 2024 analysis in PLOS Biology, the rate at which animal-tested therapies successfully translate to human regulatory approval is statistically indistinguishable from, 5%.
Despite advances in computational structural biology, organoid models, and AI-driven ligand docking, our predictive power regarding human efficacy remains equivalent to a coin flip weighted heavily toward failure.
This lack of Predictive Validity provides the necessary context for the "Cambrian Explosion" currently reshaping Chinese biotech. While US observers dismiss it as "fast-follow" copycatting or the result of loose regulation, they are missing the deeper Industrial Math.
China hasn't just deregulated; they have operationalized the only game-theoretical strategy that makes sense in a world with 5% predictive validity: High-Throughput Serendipity.
The "Target-Based" Fallacy (Sadri’s 9.4%)
The Western pharmaceutical edifice is built on the dogma of Target-Based Drug Discovery (TBDD): Validate a biological target, engineer a high-affinity ligand, and engage the mechanism. It is elegant, reductionist, and, according to historical data, mostly inefficient.
In a landmark 2023 study published in the Journal of Medicinal Chemistry, researcher Arash Sadri analyzed the origins of every small-molecule drug approved by the FDA. The results challenged the core tenets of TBDD:
Only 9.4% of approved drugs were discovered via target-based assays.
The vast majority originated from Phenotypic Screening (testing compounds in functional cellular models without knowing the mechanism) or serendipitous clinical observations.
If over 90% of pharmaceutical successes come from "messy" phenotypic luck, why does the US industry allocate the vast majority of its R&D budget to "clean" target validation? We are over-engineering the input for a system that is governed by stochastic output.
The Scannell Equation: Validity vs. Throughput
This empirical reality creates a conflict with Jack Scannell’s decision-theory framework (often cited alongside his coinage of "Eroom's Law"). Scannell argues that R&D productivity is a function of two primary variables:
Predictive Validity (V): How well does your model predict human outcome?
Throughput (T): How many experiments can you run?
Scannell’s warning was that "running 100 bad experiments is worse than running 1 good one." The US industry interpreted this as a mandate to invest billions in maximizing V (Predictive Validity).
China’s bet is the inverse: If V is asymptotically approaching zero (as Sadri’s 9.4% suggests), then T (Throughput) is the only lever left to pull.
China vs US Biotech R&D: The Industrial Math of "Implied Approval"
China is winning the "Throughput War" not by being smarter, but by altering the unit economics of experimentation.
1. Regulatory Velocity (The 60-Day Clock)
While the US FDA’s 30-day IND review is efficient, it often triggers "Clinical Holds" that pause the clock for months of Q&A. China’s NMPA has moved toward a model of "Implied Approval" for established modalities. If the regulator does not issue a rejection within 60 working days, the trial is automatically authorized. This creates a forcing function for speed that the US "Gatekeeper" model lacks.
2. The Cost of Recruitment
In the US, the "loaded cost" of a Phase 1 oncology patient can exceed $150,000. In China, centralized hospital networks and large, treatment-naive populations suppress this cost by 40-60%. When your "Cost per Shot on Goal" is halved, you can essentially run a "Portfolio Strategy" of trials for the price of a single US "Sniper Shot."
3. The Result: 7,100 Trials
This is why, in 2024, China listed 7,100 clinical trials, overtaking the US (6,000). In the China vs US Biotech R&D race, it appears China is poised to take the lead in the near future unless the US adjusts course. seems They are effectively running a country-scale phenotypic screen, putting thousands of molecules into humans (the only model with V=1.0) to see what hits.
The "Commodity" Trap
A recent analysis in Life Science Leader correctly identified "Biotech’s Big Misconception": We treat small molecules and monoclonal antibodies as "High Tech," but they are largely Commodities.
Synthesis: Standardized.
CMC: Solved.
Targeting: Known.
As demonstrated in the solar panel and lithium-ion battery sectors, China wins commodity wars. By treating clinical data generation as a manufacturing process rather than an artisanal craft, they are scaling the "commodity" of clinical evidence.
The Way Out: Clinical Abundance & The Archive
The US cannot win a race to the bottom on labor costs or regulatory looseness, nor should we, given the "negative externalities" of environmental and safety risks inherent in that model. Instead, we must lower our Transaction Costs through policy innovation.
1. The Biotech Archive (Teslo/IFP)
Ruxandra Teslo at the Institute for Progress has proposed a "Biotech Archive", utilizing bankruptcy courts (Chapter 7) to force failed companies to open-source their datasets. The industry has spent billions generating "Negative Data" (failed tox, failed efficacy). Unlocking this data would allow us to train AI models on "failure modes," doubling our predictive power without running a single new trial.
2. Clinical Trial Abundance
We must move from "Gatekeeper" to "Gardener." Australia’s Clinical Trial Notification (CTN) scheme allows ethics committees, rather than federal regulators, to approve low-risk Phase 1 trials. Adopting a similar notification track for standard modalities in the US could cut IND start-up times by months.
3. Embrace Decentralized Trials (The "No-Site" Advantage)
While China wins on the "Cost of Brick-and-Mortar" (cheaper hospital sites, centralized labor), the US has a distinct advantage in the "Cost of Digital Infrastructure." We must leverage this to bypass the inefficiencies of physical sites entirely.
a) The Math: In a traditional US trial, the "Geographic Tax" is massive. Limiting enrollment to patients within driving distance of a major academic medical center restricts the pool to <5% of the population, driving up recruitment timelines (and cash burn).
b) The Shift: Covid-19 proved that Decentralized Clinical Trials (DCTs) are feasible. As Sandeep (my co-author/colleague) witnessed while leading BARDA’s DRIVe unit, remote monitoring and digital consent don't just "work", they broaden access and improve adherence, effectively increasing the n (sample size) while decreasing the t (time).
c) The Policy Fix: The barrier isn't technology; it is contracting complexity and regulatory hesitation.
Standardization: We need simplified, standardized contracting for remote vendors (mobile phlebotomy, home nursing) so small biotechs don't drown in legal fees.
Regulatory Signal: The FDA must explicitly lead on data quality standards for home-based collection (e.g., Tasso devices). By defining the "Quality Standard" for remote data, the FDA can de-risk the decision for small biotechs to abandon the brick-and-mortar model.
4. The "Biotech Social Contract" (Kolchinsky)
Finally, we must heed Peter Kolchinsky’s argument: The US advantage is Novelty, not Efficiency. We must defend the pricing power of the first 10 years of a drug’s life, the period that funds the serendipity, while rigorously enforcing the "Social Contract" of rapid genericization thereafter.
Conclusion
The Evolutionary Explosion in China’s biotech sector is messy, ethically fraught, and ecologically expensive. But it is also mathematically coherent. China has accepted that drug discovery is a stochastic process and built an industrial machine to service that probability. If the US continues to treat clinical trials as "Verification Events" for perfect hypotheses, we will lose. We must start treating them as "Exploration Events", and build the infrastructure to run them in abundance.
