Can Sophisticated Study Designs With Regression Analyses of Observational Data Provide Causal Inferences?