Here's some stuff about infant mortality and how it's recorded:
However, the method of calculating IMR often varies widely between countries, and is based on how they define a live birth and how many premature infants are born in the country. Reporting of infant mortality rates can be inconsistent, and may be understated, depending on a nation's live birth criterion, vital registration system, and reporting practices. The reported IMR provides one statistic which reflects the standard of living in each nation. Changes in the infant mortality rate reflect social and technical capacities[clarification needed] of a nation's population. The World Health Organization (WHO) defines a live birth as any infant born demonstrating independent signs of life, including breathing, heartbeat, umbilical cord pulsation or definite movement of voluntary muscles. This definition is used in Austria, for example. The WHO definition is also used in Germany, but with one slight modification: muscle movement is not considered to be a sign of life. Many countries, however, including certain European states (e.g. France) and Japan, only count as live births cases where an infant breathes at birth, which makes their reported IMR numbers somewhat lower and increases their rates of perinatal mortality. In the Czech Republic and Bulgaria, for instance, requirements for live birth are even higher.
Although many countries have vital registration systems and certain reporting practices, there are many inaccuracies, particularly in undeveloped nations, in the statistics of the number of infants dying. Studies have shown that comparing three information sources (official registries, household surveys, and popular reporters) that the "popular death reporters" are the most accurate. Popular death reporters include midwives, gravediggers, coffin builders, priests, and others—essentially people who knew the most about the child's death. In developing nations, access to vital registries, and other government-run systems which record births and deaths, is difficult for poor families for several reasons. These struggles force stress on families[clarification needed], and make them take drastic measures[clarification needed] in unofficial death ceremonies for their deceased infants. As a result, government statistics will inaccurately reflect a nation's infant mortality rate. Popular death reporters have first-hand information, and provided this information can be collected and collated, can provide reliable data which provide a nation with accurate death counts and meaningful causes of deaths that can be measured/studied.
And
Another challenge to comparability is the practice of counting frail or premature infants who die before the normal due date as miscarriages (spontaneous abortions) or those who die during or immediately after childbirth as stillborn. Therefore, the quality of a country's documentation of perinatal mortality can matter greatly to the accuracy of its infant mortality statistics. This point is reinforced by the demographer Ansley Coale, who finds dubiously high ratios of reported stillbirths to infant deaths in Hong Kong and Japan in the first 24 hours after birth, a pattern that is consistent with the high recorded sex ratios at birth in those countries. It suggests not only that many female infants who die in the first 24 hours are misreported as stillbirths rather than infant deaths, but also that those countries do not follow WHO recommendations for the reporting of live births and infant deaths
Another seemingly paradoxical finding, is that when countries with poor medical services introduce new medical centers and services, instead of declining, the reported IMRs often increase for a time. This is mainly because improvement in access to medical care is often accompanied by improvement in the registration of births and deaths. Deaths that might have occurred in a remote or rural area, and not been reported to the government, might now be reported by the new medical personnel or facilities. Thus, even if the new health services reduce the actual IMR, the reported IMR may increase.
And
The exclusion of any high-risk infants from the denominator or numerator in reported IMRs can cause problems in making comparisons. Many countries, including the United States, Sweden and Germany, count an infant exhibiting any sign of life as alive, no matter the month of gestation or the size, but according to United States some other countries differ in these practices. All of the countries named adopted the WHO definitions in the late 1980s or early 1990s, which are used throughout the European Union.
However, in 2009, the US CDC issued a report that stated that the American rates of infant mortality were affected by the United States' high rates of premature babies compared to European countries. It also outlined the differences in reporting requirements between the United States and Europe, noting that France, the Czech Republic, Ireland, the Netherlands, and Poland do not report all live births of babies under 500 g and/or 22 weeks of gestation. However, the differences in reporting are unlikely to be the primary explanation for the United States' relatively low international ranking. Rather, the report concluded that primary reason for the United States’ higher infant mortality rate when compared with Europe was the United States’ much higher percentage of preterm births.
en.m.wikipedia.org/wiki/Infant_mortality