Statistical Manipulation and apples to oranges comparisons.
First of all, most other OECD countries don't count every birth. They often exclude any child that died within a critical period of time or are born underweight. Since most infant deaths are exactly that, this makes other countries look like their infant mortality rate is much better than the US which counts every birth like you'd expect. This also impacts the life expectancy. When you make apples to apples comparisons (and many studies have) the US infant mortality rate is actually OUTSTANDING.
However, on top of that, the US has a couple problems. One is a fair bit of violence in our inner cities, and we drive alot more miles and have more accidents, but the big one is unhealthy habits. Especially our diets. How to put this delicately? We're fatasses. We eat preposterously unhealthy diets and obesity is strongly correlated with almost every major illness, so despite having a high quality healthcare system (again, there are apples to apples studies that compare health outcomes of similar diagnoses and the US healthcare system actually scores crazy high in them), the US life expectancy isn't so great because we have such a higher initial incidence of everything from diabetes to heart disease to cancer and much more...all correlated with bad diets and obesity.