Animal tests: The first stage is to screen drugs using animal tests to reject inactive preparations. There are no formal design requirements at this stage of the study. The design and approach depend on the intuition of the investigator.

Phase 1: Studies in humans start with the phase 1 clinical trials. These are small exploratory investigations that are not comparative. They accomplish 4 objectives: (a) determining the optimal or maximum tolerated dose, MTD. (b) studying drug administration schedules and (c) studying drug toxicity, qualitative and quantitative and (d) evidence of anti-tumor activity. In determining MTD, patients are entered at a given dose schedule. If they tolerate it, they are then put on a higher dose schedule until the limit of tolerance is reached. Patients selected for phase 1 trials have disseminated disease that is not amenable to treatment. Separate trials must be carried out for adults and children because of differences in drug kinetics. Trials for solid must be separated from those of non-solid tumors. Pregnant women are excluded. The patient must have life expectancy of at least 12 weeks. Animal data is used to decide the starting dose, dose schedule, and dose escalation.

Phase 2: Phase 2 studies are small non-comparative studies that assess therapeutic activity of a drug. The objective of phase 2 trials is to screen for anti-tumor activity in advanced disease. There are 2 sub-stages: stage 2a determines whether the drug is effective; stage 2b is a follow-up to determine a precise estimate of effectiveness. Phase 2 may be disease-oriented (test one drug on a given disease) or drug oriented (test several disease on a drug). Three designs are possible: single stage, sequential, and multi-stage (enter patients in batches). The end-points of a phase 2 trial are: response or toxicity. Efficacy is measured as the proportion of the number responding to the number treated; a cut-off of 0.2 or 0.3 is used for efficacy determination. False negative and false positive errors can occur. The sources of bias are: selection bias, wait and see attitude, and selective reporting of results. Controls may be used in phase 2 trials comparing the new agent against a standard agent. Stratification may be carried out. Cross-over designs are possible. Single or combination therapies are tested.

Phase 3: This is the main stage of clinical trials. It is a comparative clinical trial that compares two drug regimens or compares a drug regimen to a placebo. Its objective is to compare the new agent against existing agents by determining relative efficacy in a comparative trial. The trial specifically aims at (a) selecting the better treatment (b) obtaining a precise measure of the effectiveness of the treatment regimen. The following measures are taken to ensure comparability: (a) The 2 groups must be similar with the exception of the treatment being assessed. This is ensured by randomization (b) The 2 groups must be treated in exactly the same way. Double blind techniques are used to prevent the possibility of investigator bias.  Phase 3 trials may be carried out for adjuvant therapies or for advanced disease. In adjuvant studies, the important end-points are remission duration (disease-free survival) but in advanced disease the end-point is survival until death. Adjuvant trials usually require a large number of subjects. A phase 3 study has three main stages: study design and protocol development, patient accrual and data collection, and follow up and analysis. Many types of specialized personnel are involved: study chairman, statistician, clinical research associates, and the data coordinator.

Phase 4: Phase 4 studies involve post-marketing surveillance by collecting data on short term and long term effects.