The information, reported by many agencies and organizations and brought together in the Refugee Nutrition Information System reports, provides an extensive set of data which can be used to draw some summary conclusions about the situation over the period October 1992 through September 1994. Since the reports are only in part from surveys, and come in as they are available rather than in any systematic way, the results must be interpreted with caution, and certainly not as if they are representative of the entire situation. Moreover, in some cases there are reports of food availability, in others malnutrition, or mortality rates; only sometimes is more than one of these available for the same place and time. The results discussed below are usually at the level of camp (or area) and time period, usually month. Three indicators chosen for the summary are food availability, expressed as kcals/caput/day, which usually refers to that supplied as food aid not including any that may be available locally (which is often rather limited or nonexistent). Second, prevalences of wasting in young children (usually under five years) are reported, mainly from surveys but sometimes from clinics, generally as a prevalence of less than -2SDs weight-for-height, or less than 80% weight for height. Thirdly, crude mortality rates are fairly widely estimated, more often from registration or "body counts", but sometimes from interview in sample surveys. In this summary, these three indicators are looked at first descriptively, showing the means and ranges; and then certain associations are briefly examined.
The range of food availability reported for nine situations is shown in Figure 13a, from 115 reports. This emphasizes the wide range of calorie availabilities reported, with an average around somewhere between 1,500 and 2,000 kcals in most situations. However, the spread of reported availabilities is very large, and significant numbers of reports are below 1,500, or even 1,000 kcals/caput/day. As noted in the box earlier, for guidance, around 2,000 kcals/caput/day approximates to the usual requirement. While it is certainly the case that some of the lower reported values are underestimated, being supplemented by local food, nonetheless in many of these the low figures do indeed mean that the food availability is far below that required for health, activity, and maintaining nutritional status.
The prevalences of wasting among young children in the refugee populations, primarily from surveys, are shown in Figure 13b. It can be seen that there are substantially more reports available (n = 298). Bearing in mind that the usual level of wasting in African populations is around 5%, sometimes rising to 10 or 15% in times of severe food crisis as in droughts, the prevalences in the figure show a shocking level of malnutrition. Indeed, these are likely to be far the highest levels of wasting in Africa ever reported. Individual cases have received quite wide recognition, both in the literature and the press - Somalia for example - but the widespread high levels of wasting are very striking. Levels ranging up to 50% are common in many of the situations - Angola, Liberia, Rwanda/Burundi, Somalia, etc. - and in Southern Sudan a number of reports have been even higher. It is significant that in Mozambique, widely recognized as a well-controlled situation for the most part, levels of wasting are around normal. The risk of mortality involved with wasting is well known, and indeed can be demonstrated from these data, as discussed below in the context of Figure 16.
Mortality rates are also quite widely reported, and require some care in the interpretation. Crude mortality rates, that is the total number of deaths per population per year, tend to be around ten per 1,000 per year, in both developed and developing countries - the similarity is due to the very different age structures between these two. This translates to around 0.3/10,000/day. The reported mortality rates as shown in Figure 13c are almost all far above this normal level, which would hardly be distinguishable from zero in the diagram. Therefore these mortality rates are overwhelmingly excess mortality, hence potentially preventable. Another way of getting a sense of scale is to consider that a normal crude mortality rate (10/1,000/year) is equivalent to 1% of the population per year. Thus at 3/10,000/day, 10% of the population would die during the course of the year; at 15/10,000/day, half the population would die in the course of a year. A number of the reported mortality rates are indeed getting up to the level at which 50% of the population would die during the year. As has been noted elsewhere, for example in Somalia and Rwanda, these are some of the highest mortality rates ever recorded. But striking also from Figure 13c is that greatly raised mortality is so common as to be almost usual among these populations.
Figure 13a. Kcal Availability Amongst Refugee and Displaced Populations in Africa.
Figure 13b. Wasting Amongst Refugee and Displaced Populations in Africa, Oct 92 - Sep 94
Figure 13c. Mortality Rates Reported Amongst Refugee and Displaced Populations in Africa, Feb 92 - Sep 94
Source: ACC/SCN (1993/4) RNISGoing back to the distribution of food supply, in Figure 14 the number of cases falling into ranges of calorie availability are shown. This is a different way of presenting the data shown in Figure 13a, and reemphasizes that at least half of the reports are of food availabilities less than 1,500 kcals/caput/day, well below adequacy, a level at which it is almost inevitable that people will be hungry and that their nutrition will deteriorate. In fact, there are significant numbers of reports of below 1,000 kcals/caput/day being available, and other information indicates that in most of these the deficit is not made up by locally available food. It should be stressed that these reports indicate major problems of logistics and accessibility, and do not imply a lack of effort to supply adequate food. The distribution also demonstrates that shortfalls in food supply are not a matter of a few hundred calories. Attempts to more closely define requirements risk obscuring the scale of the deficits that need to be met. The issue is not whether the target should be 1,900 or 2,100 kcals/caput/day, in the first instance, but whether anything near adequacy can be achieved.
For the relatively fewer cases (n = 48) in which mortality and food supplies are available for the same place and time, the association is shown in Figure 15. Here it is striking that the very high mortality rates are associated with calorie availabilities of less than 1,500, and indeed in those cases where more than 2,000 kcals/caput/day are available, the mortality rates are on average less than 1/10,000/day, and do not reach 2/10,000/day. This diagram shows the means and the ranges, and while only having relatively small numbers of data points, seems likely to be reflecting reality. It should of course be stressed that although there is obviously reason to suppose that low calorie availability contributes causally to mortality, this is not proved by these data, and the interpretation is that situations with low calorie supply are those with high mortality. Whilst it is likely that improving the food availability would contribute to reducing mortality, health and other factors no doubt play an important role, and the association should not be overinterpreted.
Figure 14. Numbers of Records and Kcals Supplied (records are individual reports from camps)
Source: ACC/SCN (1993/4) RNIS
Figure 15. Mortality Range, mean (x), and Kcals Supplied
Source: ACC/SCN (1993/4), RNIS.Two data points omitted because they are off the scale are as follows: MR of 30 and Kcal <400 and MR of 23 and Kcals 1000-1499, from Goma.
Figure 16. Wasting vs Mortality in Refugee and Displaced Populations in Africa
Regression Line Statistics.When prevalences of wasting are compared with reported mortality, as shown in Figure 16, again the expected situation is observed, but the closeness of the association is striking. The figure uses log-scales for both mortality and wasting, and the association is quite strong, as shown in the statistics under the figure. The crude mortality rate of 1/10,000/day, although elevated, is often taken as a cut-off to define a crisis situation. It is of interest that very few situations with wasting levels of less than 10% show a crisis level of mortality by this definition. In fact, it can be calculated (results not shown here) that a good degree of prediction of mortality exceeding 1 or 2/10,000/day can be obtained using cut-offs of wasting of 10 to 20% (optimum around 15%). It should be noted that the measurements here are simultaneous whereas clearly wasting is expected to predict mortality over time; however this time period is quite limited, and simultaneous measurement does not invalidate the conclusion. It does mean that wasting can be used as a reasonable estimate of mortality risk, and that situations with over 10% of wasting in young children should be taken as likely to be critical. Turning back to Figure 13b, this demonstrates that most of the situations are indeed critically in need of additional inputs.Log CMR = A + B(LogWasting)
Intercept (A) = -2.05 (t = -8.78, p<0.000)
Coefficient (B) = 0.875 (t = 10.6, p<0.000)R2 = 0.59. n = 80
Source: ACC/SCN (1993/4), RNIS.
* * *
Widespread severe malnutrition and high risk of dying are evidently common amongst the refugee and displaced populations in Sub-Saharan Africa. Whereas famine risk has receded in all other regions, in Africa it has shifted towards these populations - almost always generated by conflict - now well outnumbering those affected by famine triggered by drought.
The numbers affected, around 20 million, are rising steadily (see Figure 3) on average, with major shifts in pattern as crises erupt and recede. The malnutrition and mortality can be controlled, and demonstrably are, with external assistance. There are several recent examples, from the interventions to control the cholera and dysentery among Rwandans in Goma, to the long-run general success in protecting nutrition for Mozambican refugees. Part of the urgency is to more rapidly achieve this control, at least when the uprooted are in contact with assistance, when they have fled to relatively safe areas where agencies can operate.
Episodes like the Rwandan crisis can be seen as a possible early warning of even worse to come. Many populations in Sub-Saharan Africa are vulnerable to ethnic conflict, some of which persists for years, liable to flare-up into full-scale war. Several such populations number in the tens of millions, and even more massive refugee flows than have been seen recently could occur, bringing famine on a wide scale. The future nutritional situation in many parts of Sub-Saharan Africa may depend, more than anything, on whether these conflicts are dampened down, or whether they explode.
