Constructing Inverse Probability Weights for Marginal Structural Models

• Published in: American journal of epidemiology Vol. 168; no. 6; pp. 656 - 664
• Main Authors: Cole, Stephen R., Hernán, Miguel A.
• Format: Journal Article
• Language: English
• Published: Cary, NC Oxford University Press 15-09-2008; Oxford Publishing Limited (England)
• Subjects: Acquired Immunodeficiency Syndrome - drug therapy; Acquired Immunodeficiency Syndrome - epidemiology; Analysis. Health state; Antiretroviral Therapy, Highly Active; Bias; bias (epidemiology); Biological and medical sciences; causality; confounding factors (epidemiology); Epidemiology; Female; General aspects; HIV-1; Humans; Logistic Models; Male; Medical research; Medical sciences; Miscellaneous; Multicenter Studies as Topic; Practice of Epidemiology; Probability; probability weighting; Public health. Hygiene; Public health. Hygiene-occupational medicine; regression model; Research methodology; Statistical methods; confounding factors (epidemiology); regression model; bias (epidemiology); probability weighting; causality; Human; Probability; Confounding factor; Epidemiology; Weight; Regression model; Methodological bias; Structural model; Causality; Public health
• ISSN: 0002-9262; 1476-6256
• DOI: 10.1093/aje/kwn164

The method of inverse probability weighting (henceforth, weighting) can be used to adjust for measured confounding and selection bias under the four assumptions of consistency, exchangeability, positivity, and no misspecification of the model used to estimate weights. In recent years, several published estimates of the effect of time-varying exposures have been based on weighted estimation of the parameters of marginal structural models because, unlike standard statistical methods, weighting can appropriately adjust for measured time-varying confounders affected by prior exposure. As an example, the authors describe the last three assumptions using the change in viral load due to initiation of antiretroviral therapy among 918 human immunodeficiency virus-infected US men and women followed for a median of 5.8 years between 1996 and 2005. The authors describe possible tradeoffs that an epidemiologist may encounter when attempting to make inferences. For instance, a tradeoff between bias and precision is illustrated as a function of the extent to which confounding is controlled. Weight truncation is presented as an informal and easily implemented method to deal with these tradeoffs. Inverse probability weighting provides a powerful methodological tool that may uncover causal effects of exposures that are otherwise obscured. However, as with all methods, diagnostics and sensitivity analyses are essential for proper use.