At what point is it valid to apply a predictive statistical model that was developed from population studies to health decisions for an individual person?

April 08, 2011

• By: Douglas S. McNair, MD, PhD

This question was posed by a reader in response to my previous blog post about the anticonvulsant medication valproic acid and drug-induced liver injury that occurs in a modest percentage of patients where the risk disproportionately involves individuals with a particular gene polymorphism.Measuring or predicting risk for an individual vs. measuring or predicting risk for populations is an important distinction to make. Epidemiologists call applying a population result to a person the "atomistic fallacy." The converse (applying risk that is known for an individual or for a small group to a larger population that is different from the individual or small group) is called the "ecological fallacy."

Read full story

• Category: Personalized Medicine, Genomics, Personal Health
• Comments (0)
• Rating: 5/5
• Tags: cardiac, womens health, consumer health, genomics, personalized medicine, diagnostics, prognosis, statistics, prediction, decision support, risk, health policy