Misunderstanding Population Statistics

In today's New York Times, Nicholas Kristoff demonstrates a gross inability to adequately interpret the raw data of a population study or the multitude of variables contained therein. In his Op/Ed piece today entitled "Health Care? Ask Cuba," Kristoff cites a CDC report that provides overall infant mortality rates by country. Using the data set, Kristoff attempts to demonstrate that our overall greater infant mortality rate than countries such as Cuba and China is a searing indictment of our public health infrastructure and that

"............. for those on the bottom in America, life in our new Gilded Age is getting crueler."

But do Kristoff's raw numbers tell the whole story? Let's break down what Kristoff states and see what the numbers actually tell us.

"In every year since 1958, America's infant mortality rate improved, or at least held steady. But in 2002, it got worse: 7 babies died for each thousand live births, while that rate was 6.8 deaths the year before."

This must mean that our ability to provide prenatal and antenatal care has declined and that the ravages of the Iraq war, the epidemic of uninsured, and a Republican system of indifference has finally taken its toll. Right? I mean what else to make of this statement that Kristoff slips in:

"Bolstering public health isn't as dramatic as spending $300 million for a single F/A-22 Raptor fighter jet, but it can be a far more efficient way of protecting Americans."

Clearly Kristoff would like us to make this association but could there be another reason for the sudden increase? Well according to experts from the CDC, and a report found here in the New Scientist, there certainly is. The rise in multiple births in the US "42 per cent between 1990 and 2002 - can help explain the increase in premature and low weight babies, write Martin and CDC statistician Kenneth Kochanek." They go on:

"The increased use of assisted reproductive therapies (ART) such as in-vitro fertilization has been strongly associated with the growth in multiple gestation pregnancies and may also be associated with an increased risk of low birth weight among singletons," they write. One per cent of all births in the US in 2001 were a result of ART."

Further associated with infant mortality (leading to birth-defects/low birth weight/multiple gestations) is advanced maternal age which is positively correlated with all the risks for increasing rates of infant death as delineated above. Couples waiting till later in life (>35) to start a family, due to the pressures of career and education, lead directly to increased measured rates of fetal compromise and infant distress/demise.

Yet Kristoff insists:

"Babies are less likely to survive in America, with a health care system that we think is the best in the world, than in impoverished and autocratic Cuba."

Now, this must be true, after all the raw numbers demonstrate decreased incidences of infant deaths in Cuba. However, as Kristoff has done throughout his article, he again does not understand how raw data is obtained (or the variables that need to be assessed in order to adequately compare the 2 disparate health care systems). Put quite succinctly:

"In the United States anywhere from 30-40 percent of infants die before they are even a day old because the United States has the most intensive system of emergency intervention to keep low birth weight and premature infants alive in the world. The United States is, for example, one of only a handful countries that keeps detailed statistics on early fetal mortality the survival rate of "infants" who are born as early as the 20th week of gestation....... in the United States if an infant is born weighing only 400 grams and not breathing, a doctor will likely spend lot of time and money trying to revive that infant. If the infant does not survive -- and the mortality rate for such infants is in excess of 50 percent -- that sequence of events will be recorded as a live birth and then a death. In many countries, however, (including many European countries) such severe medical intervention would not be attempted and, moreover, regardless of whether or not it was, this would be recorded as a fetal death rather than a live birth."

Mr. Kristoff never accounts for the disparate means of reporting between the 2 countries, the advanced technology that allows for low-birth weight infants to be born and kept alive in the US, and the differences in the philosophy regarding premature and low birth weight infants in each system.

To give one final example of how population statistic can easily be skewed by factors not assessed by Kristoff let's look at maternal tobacco usage. In recent study by Salihu et Al. ( Matern Child Health J. 2003 Dec;7(4):219-27) it was estimated that about 5% of infant deaths in the United States were directly attributable to maternal smoking while pregnant, with variations by race/ethnicity. That's 5% of Mr. Kristoff startling infant mortality statistic (a large number) directly attributable to actions of the individual and not the failures of the US public health infrastructure.

Kristoff has an agenda to sell, in this particular case, it is to spin the story of increased infant mortality (and its laymen’s misinterpretations) into an indictment of the current administration's policies regarding health-care and to advocate for the (sure to fail) socialized health-care system that the Times has been so intent upon enacting ever since Clinton's failure to do so in '93. But let us take Mr. Kristoff on his word for a moment and allow for the fact that he truly believes what he has written in today's Op/Ed piece. Perhaps then, when his Grandchildren are in need of prenatal and antenatal care, we will find that Kristoff has shipped them to Havana where they can get far superior care in Castro-land as opposed to the desperately lagging US.

Or maybe not.