'Famine in slow motion': A retrospective study of North Korean mortality, 1995--1998

In complex humanitarian emergencies, two central indicators of population health are nutritional status (particularly among children under five) and mortality rates. A deteriorating economy, coupled with a series of natural disasters in 1995--1997 led to a severe food crisis in North Korea, officially known as the Democratic People's Republic of Korea (DPRK). Although the government sought and received large amounts of international food aid beginning in 1995 to alleviate a serious food crisis, restrictive government policies permitted only intermittent assessments of nutritional status among children and prohibited internationally supervised assessments of population mortality.; Lacking direct means of measuring recent mortality in North Korea, it became necessary to explore alternatives to obtain demographic data. Since 1994 or 1995, significant numbers of North Koreans had been moving across the Chinese border in search of food for themselves and their families. It was estimated that between 50,000 and 150,000 North Koreans were staying temporarily in China, principally in Yanbian Korean Autonomous Prefecture, which is home to nearly one million Korean-Chinese. In 1999--2000, interviews were conducted with a total of 2,692 North Koreans who had crossed into China. Respondents provided information on births, deaths, and migration patterns in their household in North Korea between 1995 and 1998.; The demographic survey of North Korean migrants in China sought to establish, using a framework of "generalizations" about famine proposed by Dyson and O Grada (2002), that a famine had occurred in North Korea during the period, 1995--1998. Recognizing certain biases and limitations, the study found, first of all, that mortality in North Korea had increased significantly. Relative to the baseline crude death rate from the 1993 North Korean census (5.4 per 1,000), the crude death rate in the sample population was well above normal (16.2 per 1,000) in 1995. Within the four-year period, using 1995 as the reference year, the risk of dying in 1997 essentially doubled.; Second, the study found that increased malnutrition was the principal cause of mortality and that mortality was also associated with increased infectious disease. Increased mortality risk also was associated to a significant degree with the lack of international food aid and with low levels of government food rations.; Third, the mortality rates of males increased by more than the mortality rates of females. In absolute terms, the difference between 1993 rates and 1995--1998 rates---what has been referred to as the female mortality advantage during famines---was significant: male crude death rates increased by 23.0 per 1,000 per year during the four-year interval, while female crude death rates increased by 16.4. Female age-specific death rates in the 1993 North Korean census were lower than for males at every age group. In the sample population, female age-specific death rates generally were lower than for males, with the significant difference being in the age group 0--4.; Methodologically, the results of the study suggest that it is possible to obtain useful demographic data in situations where the population of interest is, at least partly, inaccessible. Surveys of refugees and migrants from disaster-affected areas---though subject to selection factors and other potential biases---should be explored whenever it is not possible to go directly to the site of the emergency.