From Molecules to Medicines: How Drug Development Evolved and Why It’s More Complex Than Ever

Early in my career in the late 1990s at Rhone-Poulenc Rorer (one of the predecessor companies of Sanofi), I had the opportunity to lead the launch of several compounds as a young marketer.

At the time, reviewing data packages with colleagues often led to recurring questions:

• Why don’t we have this data?
• Why don’t we have that analysis?
• This would help us differentiate more clearly from competitors.

As I moved on to subsequent launches at a global level, those same questions continued to surface. The expectation remained that more data, particularly data supporting differentiation or addressing specific concerns, would translate into stronger positioning and greater impact at launch.

Only later, through involvement in medical affairs, participation in label negotiations, and subsequent work across early commercialization and portfolio strategy—as well as direct involvement in clinical development across multiple programs—did the complexity behind those “missing pieces” become clear.

Drug development is not a linear process of generating all desired data before launch. It is a long, iterative journey shaped by scientific uncertainty, operational constraints, regulatory expectations, and increasingly, payer considerations.

Today, that complexity continues to grow. New modalities, rising development costs, and heightened expectations from regulators and health technology assessment bodies mean that success is no longer defined by approval alone, but by the ability to deliver a reimbursable and differentiated therapy.

From Chemistry to Biology and Beyond

For much of the twentieth century, pharmaceutical innovation was driven by chemically synthesized small molecules. Advances in medicinal chemistry enabled the development of therapies that transformed treatment across multiple disease areas.

A major inflection point came with the rise of biotechnology. Companies such as Genentech helped establish recombinant DNA technology as a foundation for modern drug development, leading to the approval of recombinant human insulin in the early 1980s and ushering in the era of biologics.

Since then, the industry has undergone a fundamental shift. Biologics, including monoclonal antibodies and recombinant proteins, have become central to innovation, particularly in oncology and immunology. Today, biologics represent more than 40% of global pharmaceutical revenue.¹

At the same time, new therapeutic modalities, including cell and gene therapies (CGTs), as well as RNA-based medicines, are expanding the boundaries of what is scientifically and clinically possible. These advances, while transformative, introduce additional layers of complexity in development, manufacturing, and long-term evaluation.

As the industry has evolved from predominantly small-molecule drugs to a broader range of biologics and advanced therapies, the approaches to discovery and early development have also changed.

The increasing diversity of therapeutic modalities has introduced important differences in how drugs are discovered and developed. While small molecules and biologics follow similar overall development stages, the underlying approaches, evidence requirements, and technical challenges differ in meaningful ways, particularly in early development (Table 1).

Development as a Series of Decisions Under Uncertainty

While drug development is often presented as a sequence of well-defined stages, in practice it is better understood as a series of “go/no-go” decisions made under uncertainty. At each stage, teams must determine whether the available evidence is sufficient to justify continued investment.

Early decisions focus on target selection and biological rationale, while later stages assess safety, efficacy, and the potential to demonstrate meaningful clinical benefit. One of the most critical and often underappreciated challenges lie in selecting clinical endpoints.

These decisions are frequently made at a stage when the mechanism of action is not fully understood, particularly for first-in-class therapies or in diseases with high unmet need and no established development paradigm.

In such settings, teams must balance scientific hypothesis, regulatory expectations, and feasibility, often without the benefit of historical benchmarks. By the time a product reaches launch, however, understanding of the disease, the therapy, and the competitive landscape has significantly evolved.

This creates a disconnect: the questions that commercial and medical teams ask at launch are informed by a much richer context than what was available when key development decisions were made. Importantly, development programs advance not because all questions have been answered, but because confidence in the available evidence outweighs the perceived risks.

This reality explains why the data package available at launch does not always include every analysis that teams might ideally want. Development programs prioritize data required for regulatory approval, while additional evidence, such as comparative efficacy or real-world outcomes, is often generated post-approval.

Attrition, Probability of Success, and the Limits of Prediction

Attrition remains one of the defining characteristics of drug development. Of approximately 10,000 compounds evaluated during discovery, only about 250 advances to preclinical testing. Of those, a small fraction (typically 5 to 10) enter clinical trials, and ultimately only one or two achieve regulatory approval.²

Even within clinical development, success rates remain limited. Approximately 63% of programs transition from Phase I to Phase II, around 30% from Phase II to Phase III, and about 58% from Phase III to approval.³

Despite increasing availability of data and analytical tools, predicting probability of success (POS) remains inherently challenging. While numerous frameworks and models have been proposed to estimate technical POS, no universally reliable or validated approach exists.

These models are highly sensitive to assumptions, therapeutic area variability, and evolving scientific understanding. As a result, they can inform, but not replace, expert judgment.

Ultimately, success in drug development is less about predicting outcomes with certainty and more about making informed decisions under uncertainty using the best available evidence at each stage.

A Multi-Modality Development Landscape

The shift from small molecules to biologics—and now to advanced therapies—has fundamentally changed how drugs are developed. Biologics, such as monoclonal antibodies, offer high specificity and the ability to target pathways that were previously inaccessible.

However, they also introduce important considerations, including immunogenicity, distinct pharmacokinetic and pharmacodynamic profiles, as well as manufacturing complexity, scalability, and cost.

Emerging modalities, including CGTs, require even more tailored approaches. These often involve novel trial designs, small and difficult-to-recruit patient populations, and long-term follow-up to assess durability and safety.

As a result, development strategies are increasingly modality-specific. Approaches that apply to small molecules do not necessarily translate to biologics or advanced therapies, requiring greater integration across scientific, clinical, and commercial perspectives.

From “Approvable” to “Reimbursable”

Another critical evolution in drug development is the shift from achieving regulatory approval to demonstrating value in a broader healthcare context. Payers and health technology assessment bodies now expect evidence not only of efficacy and safety, but also of comparative effectiveness, real-world outcomes, and economic value.⁴

These expectations increasingly influence development strategies early in the process, shaping decisions related to endpoints, comparators, and study populations. At the same time, development decision-making is becoming increasingly cross-functional.

While clinical development and regulatory teams have traditionally led these decisions, leading organizations now incorporate earlier input from medical affairs, health economics and outcomes research (HEOR), and early commercialization teams.

Medical affairs contribute critical insights into unmet medical needs and the realities of patient care, while early commercialization teams bring perspective on the evolving treatment landscape, future competition, and how a therapy may ultimately be positioned. Together, these perspectives help shape development strategies that are not only scientifically and regulatorily sound, but also relevant to clinical practice and capable of supporting access.

Looking Ahead: Reducing Uncertainty in an Increasingly Complex System

Drug development has long been described through Paul Ehrlich’s “four G’s”: Geld (money), Geduld (patience), Geschick (skill), and Glück (luck).

While all four remain relevant, the industry continues to strive toward reducing reliance on luck through better science, improved data, and more informed decision-making.

Advances in human genetics, biomarkers, and data analytics are improving target selection and trial design. At the same time, greater integration of clinical, regulatory, and cross-functional perspectives is helping to align development strategies with real-world needs.

Even so, uncertainty remains inherent to the process. Developing new therapies will continue to require not only scientific rigor and strategic discipline, but also the resilience to navigate the unknown.

Understanding this complexity is essential, not only for those directly involved in research and development, but also for professionals across the pharmaceutical ecosystem who contribute to bringing new medicines to patients.

About the Author

Natalia Borinshteyn, MD, PhD, is Founder and President of Life Science Excellence Inc., a boutique strategic consulting firm supporting pharmaceutical, biotech, and medical device companies. She brings more than 20 years of experience across clinical development, medical affairs, and commercial strategy, and has contributed to the development and launch of multiple innovative therapies across dermatology, immunology, gastroenterology, diabetes, and HCV.Previously, Dr. Borinshteyn served as Vice President of the HCV Launch Team at AbbVie, where she led Global and US Medical Affairs and played a key role in launch readiness and execution. Earlier in her career, she held leadership roles at Sanofi across clinical development, portfolio strategy, and commercial operations.Dr. Borinshteyn holds an MD and PhD in immunology and completed executive education at The Wharton School and Harvard Business School.

REFERENCES

1. Evaluate Pharma. World Preview 2023, Outlook to 2028. Evaluate Ltd.; 2023.
2. Paul SM, Mytelka DS, Dunwiddie CT, et al. How to improve R&D productivity: the pharmaceutical industry's grand challenge. Nat Rev Drug Discov. 2010;9(3):203–214.
3. Hay M, Thomas DW, Craighead JL, Economides C, Rosenthal J. Clinical development success rates for investigational drugs. Nat Biotechnol. 2014;32(1):40–51.
4. IQVIA Institute. Global Trends in R&D 2023. IQVIA; 2023.
5. BIO Industry Analysis. Clinical Development Success Rates 2011–2020. Biotechnology Innovation Organization; 2021.
6. González Cabrera D, Selzer PM, Spangenberg T. A brief history of the pharmaceutical industry. In: Praziquantel: Discovery and Development of an Anthelmintic Drug. 2026.