Confoundings and Interactions Assessed as Dependent Adverse Effects
With confounding a subgroup performs better for all treatments in a trial. With interaction a subgroup performs better for one treatment. Both confounding and interaction are adverse effects that may obscure the overall treatment efficacy of a trial, and they are adverse effects of the dependent type, i.e., they must be significantly related not only to the intervention but also to the outcome. In this chapter examples are given, as well as methods for assessment.
Also methods for assessing the presence of dependent adverse effects in multi- instead of mono- exponential mathematical models are discussed. The multiplication of such models as required for interaction/confounding assessments, is mathematically too complex for practical use, and they can, therefore, not be adequately used for assessing their presence. Alternatives, are, however, available.