Wednesday, December 5, 2007

Wednesday Math, Vol. 6: Infant mortality and the mismeasure of rankings

A few weeks back, the topic of death statistics was discussed, and because deaths per year are relatively rare occurrences, the scale that is used is deaths per 100,000 population. Using that scale, the numbers look like whole numbers, which people like better than fractions or decimals. When we see that traffic deaths are at about 13 per 100,000 population while deaths from heart disease are at about 240 per 100,000, we can get a sense of the scale difference and it's easier to read than saying .013% and .24%. People get confused when both decimals and percentages are combined. I know this from experience in the classroom. When I teach about this, I ask the class what the legal limit for blood alcohol is, and someone will always volunteer "point oh eight", or .08, when it's actually .08%. It would be clearer if we said that 80 parts alcohol per 100,000 parts blood by volume is the limit, as it would avoid this percent AND decimal confusion.

One death statistic is kept on a different scale, and that is infant mortality, where the scale is per 1,000 live births. The definition of infant mortality is how many babies are born live but do not see their first birthday. As you might expect, in general terms, the more prosperous a country is, the lower their infant mortality rate. On the current CIA World Fact Book list of 221 countries, here are the best five for 2007.

1. Singapore: 2.30 per 1,000 live births
2. Sweden: 2.76 per 1,000 live births
3. Japan: 2.80 per 1,000 live births
4. Hong Kong: 2.94 per 1,000 live births
5. Iceland: 3.27 per 1,000 live births

The United States currently ranks 42nd, at 6.37 per 1,000. This puts us between South Korea and Croatia. We lag behind every country that can be considered "industrialized".

People like ranking systems, but they skew the data. 42nd sounds bad, and while it certainly shows there is room for improvement, there are ways to present this information that are more positive. The world average in 2007, for example, is 43.52 infant deaths per 1,000 live births. There are twelve countries in the world, mostly in sub-Saharan Africa but also in Afghanistan, where the rates are over 100 per 1,000 live births.

Infant mortality rates in the United States have improved dramatically over the last century. In the picture to the left, the good looking guy in the odd looking shirt is the gosh darn pater familias, and the funny looking bald headed kid is Matty Boy about a half century ago.

[Who am I kidding? Even bald-headed, 50 years ago I was completely adorable.]

In the 1950's, the U.S. infant mortality rate was around 40 per 1,000 live births, and I was almost one of the 40. I had a bout of bronchiolitis, the inflammation of small blood vessels in the lungs. There was no treatment then, and after some options that didn't show much improvement were tried, there is still no treatment today. It's a watch and wait situation, and isn't life threatening unless it causes a bout of pneumonia. For me, it caused a bout of pneumonia. I was fed through a tube in my foot for a few days, but I pulled through.

I don't want this thought of as "spin", but other ways of looking at the data. There are a lot of positive trends in these numbers over the years, and though we should definitely try to improve the situation, it's not as though we have rates that rival the poorest nations, even in the states where the numbers are worst. (Several Southern states and Washington DC have infant mortality rates around 11 or 12 per 1,000. Not good, but not third world level bad, either.)

But I would like to look at the numbers one more way. 6.37 per 1,000 is the same as 637 per 100,000; just multiply both the rate and the scale by 100.

Infant mortality is worse than cancer and heart disease combined in the general population.

Statistically, it's more dangerous to be a baby under one year old in the U.S. than it is to be an American soldier in Iraq for a year.

There's no Lou Dobbs or Nancy Grace on TV to beat us over the head about this, but infant mortality should be a serious national priority. Other countries have shown us that we can do a lot better. All we need is the political will.

Welcome to Bolivia! Their 2007 infant mortality rate is 50.43 per 1,000 live births. That is roughly equal to the U.S. rate 60 years ago.


Dr. Monkey Von Monkerstein said...

Your head looks huge in that photo, I'm guessing it's due to your large brain.

dguzman said...

Look at that little cutie, probably already 'splainin something to Daddy (?). BTW, Daddy's shirt is awesome.

Matty Boy said...

There is no truth to the rumour that I was the original model for Charlie Brown from Peanuts, but I will admit I could have been called "the rounded headed kid".

Jess Wundrun said...

Are there any statistics (outside of the fundamentalist whacko anti-choice tribe) that show in utero deaths due to poor health care?

I am amazed that my children are estimated to have shorter life-expectancies than me.

Splotchy said...

Nice post, and nice pic of you and your papa.

Karla said...

I think that baby deserves to be on Cute Overload!

And Dad's shirt is fetching!

I remember from a class on demographics that the high total IMR in the U.S. is hiding a big fat skew in non-white, inner-city, and low-income populations (for instance, in 2000, IMR for African American babies was 14.1). But for middle and upper class whites, the rate is actually down.

To be a poor, non-white baby in America is not such a good thing. Go Republicans! You suck!

Matty Boy said...

Hi, Jess. It's hard to find data about still births and miscarriages. It's assumed there are a lot of miscarriages that happen so early the mom doesn't even know she's pregnant. When I have found data about the U.S., we have a very strange set of three facts being true.

1. Our abortion rate is higher than much of Europe where abortions are legal, though much of red state America has very low abortion rates.

2. Our reported miscarriage rate is also fairly high for an industrialized nation. (Could these be illegal abortions in red states? I know many red state numbers are higher than blue state numbers, but the rest is speculation.)

3. Our fertility rate is definitely higher than any industrialized nation.

Go figure.

Distributorcap said...

i can see that as a baby the math problems were already being solved in your brain

Matty Boy said...

Many are the comments about my big head. I will say that it looks like my dad asked me a question, and I was taking some time to gather my thoughts and make the best response I could.

FranIAm said...

I thought I left a comment here- alas it did not take.

All I can remember is something about the cute child- cute, forget it, he is adorable.

And very math-ish.

no_slappz said...

Infant Homicide


The infant homicide rate increased from 4.3 per 100,000 in 1970 to 9.2 per 100,000 in 2000, before falling to 8.0 per 100,000 in 2004.


Homicide accounts for more than one in five injury-related deaths among infants (under one year of age) in the United States.

Infants are most likely to be killed by their mother during the first week of life but are more likely to be killed by a male (usually their father or stepfather) thereafter.

Half of all infant homicides occur by the fourth month of life, and the risk of infant homicide is highest on the day of birth. Homicide risk is greater in the first year of life than in any other year of childhood before age 17.3

Research studies of infant death data drawn from multiple agency records (such as police or social service records) indicate that the actual rate of infant deaths attributable to substantial abuse or neglect of infants and children up to four years old is more than twice as high as the official rates reported in death certificate data.

Better reporting of the circumstances surrounding infant fatalities would improve the quality of death certificate records. Studies have also indicated that a substantial but uncertain number of unreported infant homicide deaths may occur among very young infants, particularly those infants for whom no birth or death certificates are found, such as those who are born with no trained attendants and not in a clinical setting.

Key risk factors associated with infant homicides focus on the circumstances surrounding the birth of the child. Among the homicides on the first day of life, 95 percent of the victims were not born in a hospital.

Other important maternal risk factors include a second or subsequent infant born to an unmarried teenage mother (19 years of age or younger); no prenatal visit before the sixth month of pregnancy or no prenatal care; a history of maternal mental illness; a mother with 12 or fewer years of education; and premature birth (gestation of less than 28 weeks).

There is a notable absence of data on risk factors associated with males, either biological fathers or others, reflecting in part the frequency of father data missing on birth certificates.


Between 1970 and 2000, the official infant homicide rate rose dramatically - from 4.3 to 9.2 infant deaths per 100,000 children under age one. Between 2000 and 2002, the rate declined to 7.5 per 100,000, and in 2004 was 8.0 per 100,000.

Differences by Gender

In most years, males have been more likely than females to be killed during the first year of life, though in 2004 the infant homicide rate was 8.0 per 100,000 children under age one for boys, and 7.9 for girls.

Differences by Race and Hispanic Origin

Non-Hispanic black infants are substantially more at-risk than other infants of dying from homicide during their first year of life.

In 2002, the most recent year for which race-specific data are available, non-Hispanic blacks had an infant homicide rate of 18.0 per 100,000, while non-Hispanic whites had a rate of 5.1 per 100,000.

Hispanics had an infant homicide rate of 6.0 per 100,000 in 2002.