That's not how you build correlational models. You have to build a survival model and estimate a hazard ratio by access to health insurance, along with other fe

That's not how you build correlational models. You have to build a survival model and estimate a hazard ratio by access to health insurance, along with other features such as race etc. Otherwise wevcan cherry pick pairs of data points to match up with whatever belief one likes to hold. https://en.wikipedia.org/wiki/Survival_analysis https://en.wikipedia.org/wiki/Survival_analysis Even survival analysis is far from perfect because its correlational, not causal. There are other techniques such as front door criteria etc. to figure these out. https://en.wikipedia.org/wiki/The_Book_of_Why https://en.wikipedia.org/wiki/The_Book_of_Why