Simpson's paradox: an example from hospital epidemiology

Epidemiology. 2000 Jan;11(1):81-3. doi: 10.1097/00001648-200001000-00017.

Abstract

Simpson's paradox was first recognized at the beginning of the 20th century, but few examples with real data have been presented. In this paper we present an example of this phenomenon from a multicenter study on nosocomial infections, and we try to explain intuitively this type of extreme confounding.

Publication types

  • Comparative Study
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Anti-Bacterial Agents
  • Antibiotic Prophylaxis
  • Clinical Trials as Topic
  • Confounding Factors, Epidemiologic*
  • Cross Infection / epidemiology*
  • Cross Infection / prevention & control
  • Cross Infection / transmission
  • Disease Transmission, Infectious / prevention & control
  • Drug Therapy, Combination / therapeutic use
  • Female
  • Humans
  • Incidence
  • Netherlands / epidemiology
  • Prospective Studies
  • Registries / statistics & numerical data
  • Risk Factors
  • Urinary Tract Infections / epidemiology
  • Urinary Tract Infections / prevention & control
  • Urinary Tract Infections / transmission

Substances

  • Anti-Bacterial Agents