Michael A. Lawton

Sep 10, 2024 | obgyn.onlinelibrary.wiley.com | Alexandra Hunt |Laura J Bonnett |Jon Heron |Michael A. Lawton

3.2.4.2 Handling Predictors Continuous predictors are optimal when not dichotomised [19]; 38 studies with well-described handling were low-risk, whilst 25 studies, with five introducing bias through dichotomisation and 20 inadequately explaining cut-off methods, were high risk. The methods used to select predictors is extremely important.