Is Infant Mortality a Good Measure of the Quality of Health-care Systems?

From Brian Ferguson at Canadian Econoview:

Canadians are used to hearing comparisons between infant mortality rates here and in the US used as an argument for the superiority of our state-run health care system. For example, in 1998, the overall infant mortality rate for the US was 7.2 per thousand live births, compared with 5.4 per thousand in Canada...

Infant mortality comparisons are tricky things. For example, mortality risk is closely linked to birthweight: low birthweight means high mortality risk. With that in mind, let's look at some detailed figures from the same periods as those averages were calculated for. In these figures, "g" is grams, and 2500g and up would be normal birthweight.

Deaths/1000 live births in Canada vs. the U.S.:

1500 - 1999g: 28.7 in Canada, 29.0 in the U.S.
2000 - 2499g: 12.4 in Canada, 12.5 in the U.S.
> 2500g: 2.3 in Canada, 2.6 in the U.S.

The infant mortality rates in the U.S. are slightly higher than those in Canada in these weight ranges, but the difference is VERY slight (what statisticians would call statistically insignificant). The data presented by Brian Ferguson show that the differences go the other way for babies born with very low birth weights: the infant mortality rate is higher in Canada in the U.S. There are many more data available at his site, along with his usual pithy discussion and analysis; please read the whole thing before dashing off a comment here.

While there are many possible explanations for the differences, one conclusion that emerges is that rough comparisons of infant mortality rates are probably not a good indication of the overall quality of medical care.