MPH Master pieces

The numbers don't lie

Here's an interesting piece. It's one of those things that looks a bit specialised but when you think about it has enormous ramifications.

In the USA there used to be a strong racial bias in liver transplant, to the disadvantage of Black patients. But since an objective scoring system (the MELD score) was introduced in 2002 this disparity has disappeared. So the ramification is this - how many other disparities or inequities could be abolished if we paid more attention to measurement rather than subjective opinion or guesswork?

And a footnote - introducing the MELD score didn't abolish gender disparities - women were less likely to receive a transplant than men both before and after MELD. In an accompanying editorial two possibilities are suggested - firstly the spectrum of liver disease in women differs from that in men. Secondly women are smaller than men - and all allocation systems allocate small livers first to children. So sometimes a disparity reflects a biological reality not an injustice.

Politics and health

Here is an attempt to relate political conditions to health. It's rather iffy, as an accompanying critique points out.

I can't help feeling that this is sort of research which gets published if it supports the conclusion the authors wanted but quietly binned if not. We all know that data dredging - looking through large datasets for the only 'significant' result - is wrong, but there is a more subtle form: trying out a hypothesis on a whole series of datasets until one permutation reaches the desired conclusion. The problem is that you can't tell from the published conclusions how many other groupings of countries, definitions of political parties, outcomes etc etc were looked at in addition to the one combination which made it into print. One marker of sound research is very strong natural categories or groupings, and as the critique points out these are lacking - the categories used in this study are all rather artificial or imprecise.

Anyway, an instructive read if you have the time.

Sure start evaluation

This paper is a must-read: it's the second evaluation of the English Sure Start programme. Sure Start aims to provide support to parents of pre-schools children. The first evaluation was disappointing and suggested that the programme might even be doing harm - this one has much better news to bring. 

The authors suggest that results have improved because the early implementation simply allowed each programme to do its own thing; now the programme is more tightly specified and controlled.

As always read the introduction and discussion sections of the paper with care, and then dip into method and results. Because the government committed wholesale to the programme the design is a comparison of Sure Start areas with non-Sure Start areas.