Fig. 4From: Testing for carryover effects after cessation of treatments: a design approachImpact of altering inclusion criteria upon differences in cumulative diagnosis rates in a parallel-group design. The x axis indicates the length of carryoverBack to article page