Treatments with moderate but important effects

This section addresses the common situation in which treatments may differ only moderately from each other, but these differences may nevertheless be important. In these circumstances we need to consider how to get valid estimates of these differences.

In this sub-section:

A variety of study designs have been used, including:

• Comparing patients given treatments today with apparently similar patients given other treatments in the past for the same disease
• Comparing apparently similar groups of patients who happen to have received different treatments in the same time period
• Unbiased, prospective allocation to different treatments
• Ways of using unbiased (random) allocation in treatment comparisons
• Following up everyone in treatment comparisons
• Dealing with departures from allocated treatments
• Helping people to stick to allocated treatments