This seems more like correlation than causation. I would expect the people whose lifestyles allow them to regularly get good sleep are also the people who are i

This seems more like correlation than causation. I would expect the people whose lifestyles allow them to regularly get good sleep are also the people who are in a good position to avoid various health problems compared to say someone working 3 jobs to make ends meet or someone clubbing till 3 am regularly. Certainly in their data, people with less than 5 hours per night of sleep were disproportionately in lower career positions, less educated, minorities, people with lower levels of fruit consumption, very high or very low BMI, current smokers, and more likely to have hypertension. These specific factors were adjusted for, but the paper does not say how that adjustment was made or what the difference in the results is before and after the adjustment. I also found it interesting that, in the reverse of most of the other factors, a disproportionate number of those who get less than 5 hours of sleep don't consume alcohol (or at least reported 0 for their average weekly consumption). Also it's weird that those sleeping less than 5 hours and those who sleep more than 9 hours seem to share the same socio-demographic characteristics.