Triangulation, in epidemiology, refers to integrating and comparing results from several studies (or several different methods applied to the same study data) that use different approaches to test causal effects, where each approach has different (and unrelated) key sources of bias. Ideally with the different key sources of bias in different studies resulting in bias in different directions.
If results from MR ‘triangulate’ with those from other approaches (with different key sources of bias), this increases confidence that those similar results are reflective of the correct causal effect. The use of different MR methods (with different assumptions and key sources of biases) can also be valuable to increase confidence in MR results, where consistent results across methods are observed.
- Lawlor DA, Tilling K, Davey Smith G. Triangulation in aetiological epidemiology. Int J Epidemiol 2017.