Categorical Predictors Assessed as Dependent Adverse Effects

  • Ton J. Cleophas
  • Aeilko H. Zwinderman


In this chapter examples are given of clinical studies, where the adverse effect has a categorical, rather than continuous pattern. The presence of the adverse effect can be easily missed, if its categorical pattern is analyzed as a linear variable. The linearly structured variable race was such an adverse effect of the effects of age and gender on an outcome like physical strength score. Also, the linearly structured numbers of concomitant medications was such an adverse effect of the effect of age on the risk of iatrogenic admissions to hospital.

A linear or logistic regression with categories assessed as separate independent variables instead of a single continuous variable is adequate for analysis, and, in the examples given in this chapter, it was able to demonstrate the presence of dependent adverse effects in the form of significant and insignificant predictor categories on the outcome.


Dependent adverse effect Categorical data Continuous data Linear regression Categorical predictor Significant and insignificant predictor categories 

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  • Ton J. Cleophas
    • 1
  • Aeilko H. Zwinderman
    • 2
  1. 1.Albert Schweitzer HospitalDepartment MedicineSliedrechtThe Netherlands
  2. 2.Department of Biostatistics and EpidemiologyAcademic Medical CenterAmsterdamThe Netherlands

Personalised recommendations