Here’s the Heritage Foundation’s Robert Book on different methods of defining infant mortality (from a letter in the March 30th Wall Street Journal):
Your editorial “The March of Health Progress” (March 25) correctly celebrates the increase in life expectancy, decrease in infant mortality and decreasing death rates from most of the leading fatal diseases in the latest annual summary of mortality data from the Centers for Disease Control and Prevention.
However, a closer look at the data reveals an even rosier picture. Increased access to fertility treatments has resulted in more multiple births and a higher average maternal age. These factors result in more high-risk pregnancies and would be expected to increase infant mortality—yet infant mortality has declined more than enough to offset these increased risks.
Those who argue that the U.S. lags behind some other countries in infant mortality fail to take into account national differences in definitions of live birth. The U.S. complies with the World Health Organization standard, which requires registration of a live birth whenever an infant shows any sign of life outside the womb, regardless of birth weight, size or duration of gestation. Many countries restrict registration to cases in which these measures exceed certain limits, such as a birth weight of 500 to 1,000 grams, a crown-to-heel length of 25 to 35 centimeters, 22 to 28 weeks of gestation, or survival for a minimum amount of time. Since small and pre-term babies are more likely to die, standards that exclude these cases artificially decrease a country’s infant mortality rate, making its health-care system seem better than it really is. Yet U.S. infant mortality rates are competitive, despite a much more broad definition of live birth.