Health scientists often use observational data to estimate treatment effects when controlled experiments are not feasible. A limitation of observational research is non-random selection of subjects into different treatments, potentially leading to selection bias. The 2 commonly used solutions to...

Several studies suggest that, on the basis of life expectancy (LE) regressions, new pharmaceutical drugs are responsible for some of the marked gains in LE observed over the last 50 years. We critically appraise these studies. We point out several modeling issues, including disentangling the...