Dictionary

The Dictionary provides concise definitions of health economics terms, arranged alphabetically for quick reference. Use it to understand unfamiliar terminology or confirm the meaning of a specific term.

11 terms matching Confounding

C

Confounding
A distortion in the estimated association between an exposure and an outcome caused by a third variable related to both.
Confounding by Aggregation
A form of confounding arising when data analysed at group level rather than individual level mixes within-group and between-group effects together.
Confounding by Indication
Confounding in pharmacoepidemiology arising when the reason for prescribing a treatment is itself linked to a patient's underlying risk of the outcome studied.

E

Effect Modification
A situation where an exposure's effect on an outcome genuinely differs across levels of a third variable, unlike confounding, which is a bias.

I

Instrumental Variable
A variable associated with the exposure of interest but with no direct effect on the outcome except through that exposure, used to address confounding.
Instrumental Variable Analysis
A technique estimating a causal effect despite unmeasured confounding, using a variable that affects treatment but not the outcome except through treatment.
Instrumental Variable Design
A study design using a variable affecting treatment assignment but with no direct outcome effect, to estimate causal effects despite unmeasured confounding.

M

Matching Design
A study design pairing treated individuals with similar untreated individuals on shared characteristics, to reduce confounding when estimating causal effect.

P

Propensity Score MatchingPSM
A statistical technique pairing treated individuals with similar untreated individuals based on an estimated propensity score, balancing measured confounding variables.

R

Regression Adjustment
A technique using a regression model to control for measured confounding variables when estimating an exposure-outcome association.

U

Unmeasured Confounding
A situation where a true confounder exists but was not recorded, preventing statistical adjustment for its distorting effect on an association.