Instrumental variable

Health Management Glossary

Research Methods

Definition. An instrumental variable is a variable that influences the exposure of interest but affects the outcome only through that exposure, which allows causal effects to be estimated when there is unmeasured confounding.

Explanation

A valid instrument must be associated with the exposure, must not share causes with the outcome and must have no direct effect on the outcome. Angrist, Imbens and Rubin (1996) showed that the method estimates the effect among those whose exposure is changed by the instrument.

In health services research, distance to a hospital, physician preference and regional variation are common instruments, as in McClellan, McNeil and Newhouse’s (1994) study of intensive treatment for myocardial infarction. Weak or invalid instruments can produce more bias than ordinary regression.

Related concepts

Causal inference Confounding Propensity score Directed acyclic graph (DAG)

Read on The Health Topics

• The Most Popular Research Methods and Analysis Techniques in Management and Organisation Research
• DAGs and Target Trial Emulation for Oncology Management Research: Eight Causal Chains
• The Sense and Nonsense of Effect Size: What Does Funder and Ozer’s (2019) Call Mean for Health Research?
• Machine Learning in Surgical PROMs: A Critical Academic Appraisal and a Forward Research Agenda
• Structural model robustness checks in PLS-SEM

References

Angrist, J. D., Imbens, G. W., & Rubin, D. B. (1996). Identification of causal effects using instrumental variables. Journal of the American Statistical Association, 91(434), 444–455.

McClellan, M., McNeil, B. J., & Newhouse, J. P. (1994). Does more intensive treatment of acute myocardial infarction in the elderly reduce mortality? Analysis using instrumental variables. JAMA, 272(11), 859–866.

Cite this entry: Kurutkan, M. N. (2026). Instrumental variable. In The Health Topics glossary. https://healthtopic.org/glossary/instrumental-variable/

← Back to the glossary