The impact of the African aids epidemic*1 John C. Caldwell



Download 161.16 Kb.
Page1/3
Date conversion16.04.2016
Size161.16 Kb.
  1   2   3

Health Transition Review, Supplement 2 to Volume 7, 1997, 169–188



The impact of the African AIDS epidemic*1
John C. Caldwell

Health Transition Centre, Australian National University

The contemporary AIDS epidemic can be compared with the other major visitations of pestilence. In Europe 20 million or more people probably died during the Black Death in 1347-1351, and globally perhaps 20 million died during the 1917-1919 influenza epidemic. By the end of 1996 the world estimates for the AIDS epidemic were over 6 million dead and a further 23 million seropositive and nearly all certain of death. In numbers dying, the AIDS epidemic will certainly far exceed both other historic epidemics, although no-one knows the total mortality in Europe and Asia for the Black Death. The reason for the inevitability of greater mortality from AIDS is the open-endedness of the present epidemic. Both the previous epidemics just cited were over in three or four years, and it is this relatively short duration which was thought to characterize epidemics. In contrast, the first AIDS cases were identified in 1981, the result of infection mostly over the previous decade. Thus, the AIDS epidemic is already a quarter of a century old, and the level of infection is still climbing both globally and in the Third World. There is no evidence as yet about its likely duration or even whether it will become endemic in some parts of the world.

There are, however, contrasting aspects of the disease. World population is now three times its level in the early twentieth century and ten times that of the fourteenth century. Population growth rates in most of the Third World are now so high that even a huge rise in mortality may not cancel them out, and we are not yet certain that any country will experience a decline in population size because of AIDS. In comparison, the Black Death ravaged a near-stationary population, reducing Europe’s numbers by perhaps one-third. The high population growth rates of the developing world have come about because of continuing high fertility together with declining mortality which has raised life expectancy even in sub-Saharan Africa to almost 50 years. Thus, the most pessimistic projection of the present epidemic does not show the expectation of life at birth falling to as low a level as 30 years in any sub-Saharan African country, while the influenza epidemic reduced India’s expectation of life for the whole intercensal decade, 1911-1921, to 18.5 years.

Nevertheless, there is in Africa a contemporary AIDS epidemic which in its intensity and in its impact on the population can be likened to the plague. In its intensity it is quite unlike anything experienced by national populations outside sub-Saharan Africa, although some sectors of other populations, such as homosexuals in the United States, may have comparable experiences. This severe epidemic is identified in Table 1 and Map 1. The affected population is found in a long belt stretching from the Central African Republic and southern Sudan through Uganda, Rwanda, Burundi, Kenya and Tanzania to Malawi, Zambia, Zimbabwe, Botswana, South Africa and Namibia. The map, but not the table, identifies southern Sudan, which has been omitted from the table both because it is not a national population and because HIV testing is so poor as to be suggestive rather than definitive. In South Africa, KwaZulu-Natal has been disproportionately affected (and the major city in the table is not Johannesburg but Durban), while the rural populations of Kenya and Tanzania have the highest HIV levels in the western parts of those countries. Parts of central and northern Mozambique also appear to have HIV levels similar to those found in the contiguous parts of the main AIDS belt (AIDS Analysis Africa 1997:7). Some of the measures, especially rural ones, may be affected by the location of the studies. Thus the Tanzanian rural estimate comes from Mbeya, one of the worst affected parts of the country. The World Health Organization’s (1995:356) estimate for the whole country implies a rural level under 10 per cent. M. Carael (personal communication 1997) challenges the Rwanda figures and reports that Ethiopia and Djibouti probably now qualify for inclusion in the main AIDS belt. He reports that Addis Ababa now records the seroprevalence level among pregnant women as nearly 15 per cent (the rural rate reported by Health Studies Branch 1997 for Jimma, a rural area, was almost nine per cent).



Table 1
Countries with the world’s highest HIV seroprevalence levels in the general (low-risk) population (percentage of adult population)





Seroprevalence levels

Dates of research reported in 1997 publication




Capital (or major city)

Outside capital (or major city)a

Capital (major city)

Outside capital (or major city)




Dates of publication


Dates of publication










1993

1997

1993

1997







Malawi

31.6

32.8

na

11.8

1995

1995

Botswana

8.8

32.4

4.1

16.0

1995

1994

Zimbabwe

18.0

32.0

12.8

16.0

1995

1993

Zambia

24.5

27.9

16.0

12.7

1994

1994

Rwanda

33.4

25.4

9.8

na

1995

-

Swaziland

2.3

21.9

na

na

1993

-

Burundi

19.9

20.0

1.6

1.8

1992

1992

Uganda

29.5

18.5

5.0

6.5

1995

1994

South Africa

1.7

18.2

na

6.4

1994

1994

Kenya

15.0

18.1

6.3

10.3

1996

1995

Namibia

2.5

17.6

na

10.3

1996

1996

Central African Republic

7.4

16.0

8.5

6.5

1993

1993

Tanzania

11.5

13.7

10.2

15.0

1995/6

1994

Source: Health Studies Branch 1993a, 1997

Note: Mostly rural, but also including provincial towns.


The population belt shown on the map is the home of about 180 million people, or three per cent of the world’s population, but is afflicted with the majority (around 55 per cent) of the world’s HIV/AIDS. This situation may change as the epidemic spreads in Asia, or it may not if the methods of containment employed in Thailand are copied elsewhere in that continent, and if AIDS continues to intensify in the main AIDS belt of Africa. In any case, the main AIDS belt is clearly the proper locus for a contemporary study of the impact of the disease, and this book does this with a particular concentration on Uganda, Tanzania, Kenya, Zimbabwe and Malawi.

Map 1
The AIDS Belt

No other part of sub-Saharan Africa is as much affected by the epidemic. Abidjan and Bobo Dioulasso in West Africa most closely resemble the major cities of the AIDS belt, but in 1995/96 each recorded a lower seroprevalence level than any of the cities listed in Table 1; in addition, there was some evidence that rural seroprevalence in Côte d’Ivoire was well below East and Southern African levels, while no data were available for rural Burkina. In much of West Africa seroprevalence was as low as one per cent in rural populations and three per cent or less in the cities. Nevertheless, sub-Saharan Africa as a whole probably accounted for almost two-thirds of the world’s HIV/AIDS.

It is clear from Table 1 that there has been quite dramatic change within the AIDS belt over the last few years. The most unanticipated change is summarized in Table 2. In East Africa, which recorded the highest seroprevalence until 1993, the levels appear to have stabilized or fallen slightly in the cities, admittedly at the very high proportion of almost a quarter of the population. Seroprevalence is still slowly increasing elsewhere in these countries, but is only half the city level. In contrast, seroprevalence in Southern Africa has soared, overtaking East Africa. In the cities it has quadrupled to about one-quarter of the population, which may turn out to be the level of stability, while in provincial areas it has reached 16 per cent, a rural level that had not been anticipated on the basis of the earlier East African experience.

Table 2
Regional change between 1993 and 1997 seroprevalence publication (unweighted averages %)aa





East Africab

Southern Africac




Capital (or major city)

Outside capital (or major city)

Capital (or major city)

Outside capital (or major city)
















1993

24.3

9.4

6.7

8.5

1997

22.7

11.1

24.4

16.0

Change

-1.6

+1.7

+17.7

+7.5

Notes:aUnweighted averages; countries omitted from comparisons if data for both dates not available.

bUganda, Rwanda, Kenya, Tanzania, Zambia, Malawi (Central African Republic omitted on geographical grounds; Burundi omitted because no new data since 1993a).

cZimbabwe, Botswana, Namibia, South Africa, Swaziland.

Sources: Table 1.


The levels in the cities are almost unbelievably high at one-quarter of the adult population. Given that the average period from infection to death is probably less than nine years in Africa, and that the adult age span covered by most of the testing is four times as long, the implication is that, if the epidemic maintains its present level over the next few decades, the majority of city people will face death from AIDS. The truth may be more complex than this, for the city may contain a substantial, if minority, population unlikely to be infected, while many of those infected may return to rural areas. Nevertheless, cities will continue to grow because the balance of rural-urban migration will continue to be towards the urban areas; indeed epidemic deaths are already opening up the urban job market. There is no certain evidence that the rural AIDS epidemic is self-sustaining rather than depending on continued reinfection from urban areas. In the Masaka longitudinal study in Uganda a high proportion of the persons identified for the first time as being seropositive had recently returned to the area, mostly from the towns (Nakiyingi 1995; Nunn et al. 1995). The experience of the TANESA project is that seropositive persons are unusually highly mobile and were probably also so before infection (Boerma et al.). In a 1990 study in Rakai, Uganda, it was shown that not only were rural HIV levels lower than urban ones, but those of small rural agricultural villages were much lower than those of trading centres, villages, or small towns with a bus stop, a few shops and access to commercial sex (Health Studies Branch 1993b, Uganda section). This pattern looks like diffusion and the continued growth of the epidemic, but in Uganda and elsewhere it could be a stable situation with infection continuing to trickle from the larger to the smaller residential areas. The Ugandan situation is repeated in Tanzania where the TANESA project found seroprevalence levels of only 4.2 per cent in genuinely rural villages but ones of 10.3 per cent in the more commercial centres along the main road (Urassa et al.). The position is somewhat blurred by Boerma et al. reporting little difference in the death rates of the two populations, but this may be a time-lag phenomenon. Alternatively, AIDS may have negated a pre-existing mortality differential in that earlier the more accessible roadside population had benefited more from medical and public health services, as well as social change, and so had lower death rates.

The reason that the sub-Saharan African epidemic has such a potential for changing the whole of society is that it is the world’s only almost exclusively heterosexual epidemic. While homosexuality, bisexuality and intravenous drug use are estimated to be the causes of 87 per cent of HIV infection in the United States, 80 per cent in Europe, and 65 per cent in Latin America, they account for no more than one per cent of the sub-Saharan African epidemic (Mann et al. 1992:117ff; Center for International Research 1994:7). In sub-Saharan African anal sexual intercourse is suppressed as being associated with witchcraft, and, while drug use is common, injected drugs are too expensive for nearly everyone. At least as many women as men are seropositive, and this also means massive vertical transmission to infants. In terms of heterosexual transmission, there are low levels of infection among girls from about 12 years of age, and among boys from some years later. The peak ages at death are in the twenties and thirties, thus removing the families’ major earners and orphaning many young children. This is aggravated by the fact that, if one parent is infected, the other is also likely to become seropositive. The African epidemic is making a major assault on the family and society, and this has been the theme of this book. As Gregson and colleagues report, there will be all kinds of demographic ramifications. These will now be addressed.



Mortality

The obvious impact of the disease, and the reason that it is so feared, is that nearly everyone who is infected is doomed to die. There is a period of grace, or the latency period, but, although this postpones mortality, it also condemns those who know their condition to years of being figuratively on Death Row. Two modifying points should be made with regard to sub-Saharan Africa. The first is that few people know they are infected until they are in the final symptomatic stage, perhaps less than five per cent. Some never know because they die as the first symptoms develop. This may lead to greater peace of mind, and is a situation apparently preferred by most of the population, but it radically reduces the possibility of intervening to reduce the level of transmission. The second is that the duration of the latency and symptomatic periods is closer to the situation in the West of the 1980s   that is, before large-scale use of medication   than was once feared. Anzala and colleagues (1995) concluded that the latency period among a sample of Nairobi prostitutes was around four years. Modelling of the early Masaka longitudinal data, noted in an Annual Report, suggested a latency period of only 4.5 years, but, with more years of survival evidence, the estimate has now been raised to seven or eight years. This may seem surprising given the other assaults on the African immune system. Similarly, the early estimate of a symptomatic period of 4.5 months, compared with 12 months in the West in the 1980s, has now been superseded by a new figure of 8-9 months. This is supported in this book by a figure of 8.6 months from the TANESA project in Tanzania (Boerma et al.). If the latency and symptomatic periods had been atypically short, this would have meant an unusually high death rate for a given prevalence level compared with experience elsewhere.

Not everyone has the same chance of infection. The situation is radically different from that of bubonic plague, influenza or smallpox. Indeed the sexually abstinent or those in what has come to be called a monogamous relationship (a misnomer because it is a description of a type of marriage and not a sexual regime) have no chance at all of being infected. We have already noted that it is more an urban than a rural disease. This book presents other findings on differentials. Carpenter et al. show the greater risk to the young, with half the seropositive females in the Masaka study being infected before reaching 25 years of age. Marriage, in spite of the fact that most wives cannot control their husbands’ extramarital sexual adventures, provides some protection. Carpenter and colleagues report of the Ugandan women in the Masaka study that, although the great majority were married, wives accounted for only 53 per cent of the seropositive, compared with 35 per cent of those never married, 12 per cent of the divorced, and four per cent of the widowed. Indeed the divorced exhibited well over double the seroprevalence level of the whole group, perhaps because some resorted to commercial sex for support. Among pregnant women at antenatal clinics in Manicaland, Zimbabwe, single women were three times as likely to be seropositive as married women, and divorced women were five times as likely (Gregson et al. 1996:13ff.). These figures are, however, subject to a selection effect, for some married women became widows because their husbands died of AIDS, while others were separated or divorced because either they or their husbands had AIDS. In all these cases the women themselves were likely to be seropositive.

The implied AIDS mortality in the main AIDS belt is staggering. Blacker and Zaba show that in Kenya, even present HIV levels mean that one-third of those reaching adulthood in this population will die of AIDS. They show that the chance of dying of AIDS is related to both the seroprevalence level and the life expectancy prior to the epidemic. Kenya, with a life expectancy of 54 years, compared with an East African average of 47 years (United Nations 1966), has relatively low mortality from other causes at each age. Where seroprevalence is higher, in the Honde Valley of Zimbabwe, Gregson et al. (1996:13ff) calculate that 50 per cent of each cohort of women will die of AIDS before 50 years of age.

Such high mortality is also recorded by the rare vital registration systems. In the Honde Valley registered deaths have doubled since 1996 (Gregson et al. 1996:17ff), and the earlier decline in infant mortality has now been reversed. These mortality and HIV levels and trends will, if unchanged, result by 2006 in male and female life expectancies of 30 and 32 years respectively instead of the 55 and 59 years anticipated in the absence of AIDS (Gregson et al. 1996:32). Change before 2006 is hardly to be expected because the period of a decade covered by the projection is little longer than the latency period of those just infected when the projection was constructed.

There must remain a possibility that Manicaland is worse affected than Zimbabwe as a whole, and that Zimbabwe’s experience will not be identical with that of Southern Africa as a whole. This may be a slim hope in the case of Botswana. Nevertheless, in the Mwanza district of Tanzania, Boerma and colleagues record a rise in mortality of one-third in the TANESA project area. It is greater in the high-risk age groups, and the mortality levels they report from males 35-44 and females 25-34 are those which might be anticipated in populations with a life expectancy of around 33 years (Coale and Demeny 1966, North Model Life Tables). The rise in morbidity and mortality may all be ascribable to AIDS even if the cause of death is recorded otherwise. Glynn et al. record that between 1986 and 1994 the incidence of tuberculosis in rural Malawi doubled because the HIV-positive were seven times as likely to develop tuberculosis as the HIV-negative.

It should be noted that the seroprevalence data in Table 1 are not nationally representative. The figures for the cities are likely to be closer to the truth than those for the rest of the country which may represent only one or two areas. Of the eleven countries in the table with new seroprevalence data since 1993 for general populations outside the major city, only three carried out national studies, two more chose 4-6 districts across the country, one chose two districts, while for five countries the information is from a single district or location. Mortality data are no better, for vital registration is almost non-existent in sub-Saharan Africa. For national data we depend on surveys which may suffer distortions in the case of AIDS deaths because the epidemic may also have killed the relative who acts as an informant or have resulted in the break-up of the family or household which serves as the information-supplying unit in such studies. As a result there is probably a substantial underestimate of under-five years mortality in AIDS-affected sub-Saharan Africa (Bicego et al. 1997:54-57) and it appears that child mortality is now rising in Zambia, Zimbabwe, Kenya and Namibia. Timaeus and Nunn are more sanguine about the estimation of adult mortality, and conclude that the orphanhood method can be adjusted to provide reasonably accurate adult figures even in the AIDS belt.

The TANESA project has provided further information on how AIDS is interpreted and treated. Without being questioned specifically on the matter, 46 per cent of persons with symptomatic AIDS and 32 per cent of those suffering from other illnesses blamed witchcraft (Boerma et al.). Doubtless many of those who did not volunteer this information also held the same belief. This helps explain why three-quarters of all AIDS sufferers sought help from traditional healers, who specialize in the occult as well as in organic malfunctions, compared with a lower proportion of those with other disorders. Nevertheless, before death half had also visited hospitals, although only one-eighth died there. Experience from elsewhere in Africa suggests that many of these visits were made not because the patients anticipated the underlying cause would be eradicated but because they wanted relief from distressing symptoms and pain. The attempt to secure such alleviation impoverishes AIDS victims’ families throughout the continent as they buy ever more medicines and analgesics. The proportion visiting hospitals is too low and varies too much through the region, to allow any estimate of seroprevalence levels, although hospitals are often the first institutions to warn of the arrival of the epidemic.


  1   2   3


The database is protected by copyright ©essaydocs.org 2016
send message

    Main page