Results 1 to 1 of 1

Thread: [Biostatistics] Confounding

  1. #1
    PharmD Year 1 TomHsiung's Avatar
    Join Date
    Jan 2013
    Location
    Chengdu, Sichuan, China
    Posts
    593

    Default [Biostatistics] Confounding

    The confounding variable, the confounder, confuses or confounds the relationship between the exposure and outcome. To be a confounder, a variable:


    • Must be associated (causally or not) with the exposure of interest.
    • Must be causally related to the outcome.
    • Must NOT be a part of the exposure-outcome causal pathway.


    When we say that a variable is associated with another variable, we mean that both variables tend to increase or decrease together. We are not saying that increases (or decreases) in one variable cause increases or decreases in the other variable directly. However, when we say a variable is causally related to another variable, we mean that changes (increases or decreases) in one variable do cause cahnges in the other variable.

    [Biostatistics] Confounding-screen-shot-2016-02-27-at-4-26-33-pm-png

    There are a number of ways of dealing with the potential for confounding among the variables (we can refer to it as controlling for confounding). Some methods we can pursue at the design stage and some at the data analysis stage. At the design stage, we can use:
    1. restriction
    2. matching
    3. randomisation;


    and at the data analysis stage, we can use:
    1. stratification
    2. adjustment.
    Last edited by TomHsiung; Sat 27th February '16 at 4:27pm.
    B.S. Pharm, West China School of Pharmacy, Class of 2007, Health System Pharmacist, RPh. Hematology, Infectious Disease.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •