Testing a pharmaceutical drug is a multi-year process that moves through laboratory experiments, animal studies, three phases of human clinical trials, and regulatory review before a drug reaches pharmacy shelves. On average, only about 14.3% of drugs that enter the first phase of human testing ever receive FDA approval, and the entire journey from discovery to approval costs an estimated $879 million when accounting for failed attempts and the cost of capital tied up during development.
Preclinical Testing: Lab and Animal Studies
Before a drug is tested in a single human being, it must prove itself in the laboratory. This stage, called preclinical testing, uses two main approaches: cell-based experiments (in vitro) and animal studies (in vivo).
Cell-based testing starts with growing human cells in a dish, typically in flat, two-dimensional layers. Researchers expose these cells to the drug candidate and observe whether it kills disease cells, slows their growth, or causes toxic effects on healthy cells. For cancer drugs, panels of dozens of cell lines representing different tumor types are used for high-throughput screening, allowing scientists to rapidly test a compound against many cancers at once. One study used 77 colorectal cancer cell lines to identify which genetic subtypes responded to a specific chemotherapy compound. More advanced three-dimensional cell cultures, which better mimic how cells behave inside the body, sometimes reveal drug effects that flat cultures miss entirely.
Animal studies come next. Mice and rats are the most common models, though zebrafish, pigs, and nonhuman primates are also used depending on the drug and the disease. These animals are chosen because their organ sizes, body functions, or disease characteristics resemble those of humans closely enough to provide meaningful safety data. The goal is to identify obvious toxicity, understand how the drug is absorbed and broken down in a living system, and get an initial read on whether it works.
Filing an IND Application
To move from animal testing into human trials, a drug company must submit an Investigational New Drug (IND) application to the FDA. This filing covers three areas: the preclinical safety and toxicity data from animal and lab studies, detailed manufacturing information showing the company can produce consistent batches of the drug, and clinical protocols describing exactly how the proposed human trials will be designed, including who will run them and how participants will be protected. The FDA then has 30 calendar days to review the application. If no safety concerns are raised in that window, human testing can begin.
Phase 1: Finding a Safe Dose
Phase 1 trials are the first time a drug enters a human body. These studies typically enroll a small number of healthy volunteers, often around 30 people, though the exact number varies. The primary goal is safety, not whether the drug works. Researchers start with a very low dose and gradually increase it, watching carefully for side effects at each level. This process, called dose escalation, aims to find the maximum tolerated dose: the highest amount that can be given without causing unacceptable harm.
A common approach is the “3+3” design, where groups of three people receive a dose, and escalation continues only if no serious toxic reactions occur. The target is typically to find a dose where no more than about 20% of participants experience a dose-limiting toxicity. In cancer drug development, Phase 1 trials are an exception to the healthy-volunteer rule. Because cancer treatments tend to be highly toxic, even initial testing is done in patients who have the disease.
Phase 2: Does the Drug Work?
Phase 2 trials shift the focus from safety to effectiveness. These studies enroll patients who actually have the condition the drug is designed to treat, and they measure whether the drug produces a meaningful response. A typical Phase 2 trial uses a single-arm design, meaning everyone receives the experimental drug and results are compared against historical data from previous treatments rather than a placebo group tested at the same time.
Patient populations in Phase 2 often include multiple subgroups with different prognoses. A trial for a cancer drug, for example, might include patients with varying genetic profiles or disease stages. Researchers must carefully account for the proportion of each subgroup, because a drug that appears effective overall might actually only be helping one subset of patients. The response rate, essentially the percentage of patients whose disease improves, is the key metric. If the response rate exceeds a predetermined threshold based on existing treatments, the drug advances to Phase 3.
Phase 3: Large-Scale Proof
Phase 3 is where a drug must definitively prove it works better than, or at least as well as, the current standard treatment. These trials are randomized and controlled: patients are randomly assigned to receive either the experimental drug or the existing treatment (or a placebo), and neither the patients nor the doctors typically know which group is which.
The scale increases dramatically. A typical Phase 3 trial might enroll 600 or more patients, and the primary endpoint is often overall survival, meaning researchers track how long patients live. The statistical bar is high. Trials are designed to have a 90% chance of detecting a real treatment benefit while keeping the probability of a false positive result at just 2.5%. When multiple experimental doses or combinations are being tested against a single control group, the statistical threshold is adjusted to be even stricter. These trials generate the core evidence that regulators use to decide whether a drug should be approved.
Quality Control Testing in Manufacturing
Alongside clinical trials, the drug itself must be tested as a physical and chemical product. This is quality control testing, and it ensures that every batch of a drug contains the right amount of active ingredient, is free of contaminants, and will dissolve properly in your body.
Dissolution testing checks how quickly a tablet or capsule releases its active ingredient. In a standard setup following United States Pharmacopoeia guidelines, six tablets are placed in baskets containing 900 milliliters of fluid that mimics stomach acid, held at body temperature (37°C), and stirred by paddles spinning at 50 revolutions per minute. Samples are drawn at intervals over roughly two and a half hours to track how the drug dissolves over time.
Purity and potency are measured using a technique called high-performance liquid chromatography (HPLC), which separates a drug sample into its individual chemical components. The instrument pumps the dissolved drug through a specialized column, and a detector identifies each component by how it absorbs ultraviolet light. This reveals whether the active ingredient is present in the correct amount and whether any unwanted impurities have crept in during manufacturing. The method must also pass robustness tests, proving it gives consistent results even when minor variables like temperature or instrument settings shift slightly.
Testing Generic Drugs
Generic drugs don’t repeat the full clinical trial process. Instead, they must prove bioequivalence, meaning the generic version delivers the same active ingredient to the same place in your body, at the same rate and in the same amount, as the brand-name drug. The FDA requires that two key measurements from blood tests fall within 80% to 125% of the brand-name drug’s values: the peak concentration of the drug in the bloodstream, and the total drug exposure over time.
This is typically demonstrated through two studies in human volunteers. One is conducted on an empty stomach, which is the most sensitive way to detect differences between formulations. The other is done after a high-fat, high-calorie meal, to confirm the generic performs the same way when food is present. Each volunteer takes both the generic and brand-name versions in a crossover design, serving as their own comparison point. If the generic meets the 80% to 125% confidence interval, it’s considered interchangeable with the original.
Phase 4: Testing After Approval
Drug testing doesn’t stop once a medication hits the market. Phase 4 studies are post-marketing requirements that the FDA can impose as a condition of approval. These studies monitor for safety signals that might not have appeared in smaller, shorter clinical trials. Some focus on a specific known side effect, tracking its frequency and severity in real-world use. Others cast a wider net, conducting broad surveillance for any unexpected adverse reactions across a defined patient population.
Phase 4 studies also fill gaps left by the original trials. Clinical trials often underrepresent certain groups, so post-approval studies may specifically examine how the drug performs in children, infants, elderly patients, pregnant women, or people of particular racial or ethnic backgrounds. Long-term efficacy studies track whether the drug continues to work over years rather than the months typically covered by Phase 3 trials. These studies have identified serious safety issues that led to drugs being pulled from the market years after their initial approval.
The Full Cost and Timeline
Developing a single new drug costs a median of about $173 million in direct expenses for the successful compound alone. But that figure is misleading, because most drugs fail. When you factor in the cost of all the failed candidates a company pursues alongside its eventual success, the figure rises to $516 million. Add in the cost of capital, the money tied up for years that could have been invested elsewhere, and the total reaches approximately $879 million per approved drug. Costs vary widely by therapeutic area, from around $73 million in direct costs for drugs treating urinary conditions to nearly $300 million for pain medications.
The overall success rate across 18 major pharmaceutical companies from 2006 to 2022 averaged 14.3%, with individual companies ranging from 8% to 23%. That means for roughly every seven drugs that enter Phase 1 human testing, only one will eventually receive FDA approval. The companies with the highest success rates aren’t necessarily the biggest; their advantage tends to come from more disciplined selection of which compounds to advance and which to abandon early.