Here is a curiously unsatisfying paper on mortality trends in the USA over the past 50 years. The great McGinnis puts his finger on the problem - mortality stats really don't tell us what we need to know in this day and age.
When I read this paper, I was reminded of Illingworth's dictum:
'Children most need love when they are least loveable.'
It's so sad - this paper confirms other studies which show that even in kindergarten, small children are heading for success or delinquency. Ok, so this is observational and doesn't gives us cause-and-effect. But you can't help feeling that what's needed is plenty of love, probably in the form of plenty of Quality Time from loving adults. In some parts of society there just aren't enough adults with time to give.
What is a 'propensity matching''? As this paper explains, it's a statistical technique for getting comparability between groups in an observational study. You wanted to do a randomised trial, allocating patients at random to one group or the other, but you couldn't: so now you have to take what you've got and use a different method to create two groups which are more or less like, or better still pair off each person in one group with someone of similar 'propensity' in the other group.
I came across this in my day job when we were trying to assess some effect (can't quite remember what) of liver transplant. You can't allocate people at random to receive a transplant or not; but also the sicker patients tend to get transplanted quicker. So comparing those who got a transplant with those who didn't is hopelessly biassed; propensity matching helps.
If you've lived a sheltered life, and I definitely have, this paper has some things to say. One standout concern for the women is that they don't trust the prison doctors.
I think (hope) that prison health care in England is much better than it was 20 years ago. Then, health care in prisons was the responsibility of the Home Office, with its own doctors and its own career structure: a small world, insulated from the NHS, with no obvious attraction to anyone who could succeed in any other branch of medicine. But responsibility was taken away from the Home Office and given to the NHS (no doubt with a lot of kicking and screaming from the Home Office). Now people in prison are looked after by regular GPs and their teams, with the added bonus of beefed up mental health care services. That's got to be a good thing.