To the Editor: We would like to respond to some points raised by Drs D’Agostino and D’Agostino1 in their Editorial that accompanied our study.2
First, our claim that our instrumental variable estimates are unbiased and approximate the average effects from randomized controlled trials (RCTs) is based on the instrumental variable being valid and highly predictive of regional cardiac catheterization rates (29%-83%).3 To be valid, an instrumental variable must have no effect on outcomes other than through its effect on treatment.4 Our instrumental variable (regional cardiac catheterization rate) was unrelated to measured patient health status. Any effect on mortality was through health system factors related to cardiac catheterization, such as higher-volume hospitals with specialized staff and equipment and coronary care units.
Stukel TA, Fisher ES, Wennberg DE. Using Observational Data to Estimate Treatment Effects. JAMA. 2007;297(19):2078–2079. doi:10.1001/jama.297.19.2078-b
Customize your JAMA Network experience by selecting one or more topics from the list below.