Clinical trial design software is software that helps a drug company plan a clinical trial before it starts, by simulating the trial many times on a computer, and then analyse the results when it ends. Planning a trial means deciding how many patients to enrol, how to compare the groups and when to look at the results. The people who do this are biostatisticians, statisticians who specialise in medicine.
Why simulate? A trial that enrols too few patients may miss a real effect, and the medicine is dropped by mistake. A trial that enrols too many wastes years and money, and gives a placebo to more people than necessary. The chance that a trial will detect an effect that is really there is called its power. Most trials aim for a power of 80% or 90%.
Take, for example, a drug company that is planning a trial of a new medicine for heart failure. The team expects the medicine to cut hospital admissions by 20%, compared with a placebo. The software runs the trial 10,000 times on the computer, with imaginary patients whose results follow the team's assumptions. It tries 1,200, 1,500 and 1,800 patients. With 1,200, only 74% of the simulated trials show the effect. With 1,800, 90% do. So the team chooses 1,800 patients.
The software can also test an adaptive design. This is a trial whose plan allows set changes partway through. Here, an independent committee looks at the results once half the patients have finished. If the medicine is clearly working, or clearly failing, the trial can stop early. The simulation shows how often each outcome would happen, so the regulator can see that the early look does not bend the result.
The expected dose often comes from in silico drug modelling, which predicts how the body handles the medicine. Some trials also use a synthetic control arm, a comparison group built from earlier patients' records, and the software works out how to compare with it.
When the trial ends, the locked data arrives from the clinical trial data platform. The statisticians run the analysis exactly as the plan said, often in general statistical software as well. A contract research organisation (CRO) usually does this work on the drug company's behalf. Much of the world's trial statistics, called biostatistics, is done by teams in India, in cities such as Pune, Bengaluru and Hyderabad.