In reply
As Williams and colleagues note, the lack of clinical data and information about general health indicators increased the likelihood of residual confounding in our analysis and was a clear limitation of our population-based study.1 We undertook the sensitivity analysis of “exclusive providers” to address the possibility that physicians selectively prescribed ranibizumab to patients with higher socioeconomic status. Although the analysis addressed the potential for selection bias within providers, it could not address confounding by socioeconomic status between providers. We openly acknowledged this limitation and the loss of statistical power associated with the analysis.