Share link: authors.elsevier.com/a/1l5tbgeXDy-p

We apply our estimator to study the effect of a reduction in circulating sclerostin levels to predict the safety and efficacy of romosozumab. We also estimate the effect of a reduction in LDL-c and waist-to-hip ratio on cardiovascular outcomes.

In realistic MR simulations, we find that our estimator is competitive with other semi- and non-parametric methods. We also identify instances where violations of statistical assumptions lead to substantial bias in MR estimates that can be avoided with Quantile IV.