Are ulcer healing and prophylaxis studies appropriately analysed?

To discuss statistical problems related to common study objectives and designs in ulcer healing and prophylaxis studies.

The nature of the study objective determines when to use a statistical analysis according to the All Patients Treated (APT) approach and when to use the Per Protocol (PP) approach. How patients discontinuing the study are handled in the statistical analysis depends on whether the discontinuation is judged to be treatment-related.

Survival analysis, in the form of life-tables, should be used more in the analysis of time to relapse. Studies with a drug-free or a reduced-dose follow-up are very often analysed incorrectly - the analysis should be based on the original randomization and not on patients healed in the acute phase.