Phase I trials focus on safety, dosage range and side effects of the new drug or treatment and involve a small number (20 to 100) of healthy participants or participants with the disease/condition.
Phase II trials focus on efficacy, but also further test safety. Participants (up to several hundred) who have the disease are enrolled and often randomly allocated to control and treatment groups.
Phase III trials often test efficacy and safety of an intervention on a large number of participants (300 to 3,000).
Phase IV trials are observational studies conducted to gather information on the effectiveness and safety of the drug or intervention in routine practice (post-marketing). They complement pre-marketing efficacy trials and can identify adverse (and occasionally desirable) effects of the treatment.
Trials conducted in well-controlled conditions, in order to evaluate the efficacy and safety of an intervention, often employing one of the following designs:
Parallel trial: A design in which participants are randomized to only one of the treatment or placebo groups and remain in that group throughout the duration of the trial. Differences in the occurrence of the study outcome provide the information needed to estimate the efficacy of the intervention.
Crossover trial: A design in which all participants receive the treatment but in a different order. For instance, participants receiving treatment A may cross over to receiving treatment B later on during the same trial and vice versa.
Delayed-Start trial (also known as randomized-start and one-way crossover trial): A design in which participants are randomized to receive either active treatment or placebo. But then, after a pre-specified period, the placebo group are switched to receive the active treatment. Often used in Parkinson's disease and dementia trials to determine whether the treatment can change the course of illness (disease-modifying effects) rather than just improve the symptoms.
Stepped wedge trial: a generalization of the delayed-start design where the active treatment is sequentially rolled-out to all participants over time. Participants are randomized with respect to the order of receiving the treatment.
Factorial trial: A design that tests the effects of more than one treatment focusing on possible interactions between them, permitting the simultaneous test of two or more different hypotheses. For instance, a study a 2X2 factorial trial can be used to assess the effects of receiving Drug A, Drug B, both drugs compared to receiving neither drug.
Trials in which groups of subjects (clusters such as household, health units, villages etc) rather than subjects are randomized to receiving the intervention. All members of a cluster receive the same intervention. Often used to study the efficacy of interventions where the effects of treatment cannot be limited to a single subject (e.g., effect of healthcare provider training on their patient outcomes) or if the effect is at least partially mediated through group effects (e.g., "herd immunity" effects in vaccine trials).
A flexible design that allows planned modifications to the clinical trial procedures (e.g., drug dose) or statistical procedures in order to improve the chance of detecting a treatment effect or reduce potential harms.
A design in which outcomes observed prior to the introduction of a new treatment are compared to that with the new treatment in similar participants. This is typically conducted when no parallel controls are available.
Trials conducted in "real-life" routine clinical practice conditions, in order to evaluate the effectiveness of an intervention.