SUMMARY CONTENTS:

This posting contains excerpts, including the preface, executive
summary and introduction, of the 54-page report released today by
the Medical Research Council South Africa, on AIDS mortality. The
full report, including figures and tables, is available, in PDF
format, at: http://www.mrc.co.za/bod

Release of the report was delayed due to controversy within the
South African government over its results, which show a rapid
increase in mortality rates among young men and women in South
Africa in recent years, and link this new age-specific increase to
AIDS. The report's authors estimate that approximately 40% of adult
deaths aged 15-49 and about 20% of all adult deaths in the year
2000 were due to HIV/AIDS.

The report's projections show that, "without treatment to prevent
AIDS, the number of AIDS deaths can be expected to grow, within the
next ten years, to more than double the number of deaths due to all
other causes, resulting in 5 to 7 million cumulative AIDS deaths in
South Africa by 2010.

+++++++++++++++++end profile++++++++++++++++++++++++++++++

Medical Research Council of South Africa, PO Box 19070, 7505
Tygerberg, South Africa; HO Tel +27 (0)21 9380911 /
Fax +27 (0)21 9380200

The impact of HIV/ AIDS on adult mortality in South Africa

Technical Report

Burden of Disease Research Unit Medical Research Council

By Rob Dorrington, David Bourne, Debbie Bradshaw, Ria Laubscher
Ian M. Timaeus

September 2001

Rob Dorrington, Centre for Actuarial Research, University of Cape
Town; David Bourne, Department of Public Health, University of Cape
Town; Debbie Bradshaw*, Burden of Disease Research Unit, South
African Medical Research Council; Ria Laubscher, Biostatistics
Unit, South African Medical Research Council; Ian M. Timaeus,
Centre for Population Studies, London School of Hygiene and
Tropical Medicine

*Correspondence to be addressed to: Dr D Bradshaw Burden of
Disease Research Unit, Medical Research Council, PO Box 19070,
Tygerberg 7505, South Africa

-------------------------------------------------------------

Preface

Malegepuru William Makgoba
President of the MRC South Africa

July 2001

In 1982, in Oxford, Dr Harold Jaffe, a senior investigator from the
Centers for Disease Control (CDC) in Atlanta presented a cluster of
cases of homosexual men who were engaged in risky sexual behaviour,
who had all the features of Acquired Immune Deficiency Syndrome
(AIDS). At this stage the Human Immunodeficiency Virus (HIV) had
not been isolated or identified yet but AIDS, as a syndrome, had
been described a year earlier by the CDC. At this meeting I
remarked that "This syndrome may be more common in Africa than it
is appreciated". I made this premature, but predictive, remark for
the following reasons: as a medical student at King Edward Hospital
in the years 1973 to 1976 I had seen several young, male patients
with Kaposi's Sarcoma and I knew homosexual behaviour was being
practised within African communities but always denied or
suppressed. In rural Sekhukhune, for example, we always heard of
the practice of 'matanyola' (sexual practice between men) and we
also heard of men who engaged in this practice, particularly in
prisons. In KwaZulu-Natal I also came to know of 'isitabane', a
Zulu word for homosexual practice.

However, when AIDS was first wrongly linked to homosexual practice
many Africans promoted the notion that homosexual practices were
'unAfrican', thus sowing the seeds for denial to justify why AIDS
would not be prevalent in their communities. This denial
predictably became the first African public response to AIDS and
swept across the continent as country after country became engulfed
in the HIV/AIDS epidemic. Today, despite many documented cases of
homosexual practice in Africa, this denial continues. The AIDS
denial was later compounded by stigmatisation, chauvinism, the
distortion of scientific information and ignorance.

In 1985, actor Rock Hudson died of AIDS. Much later Freddie Mercury
of Queen and Rudolf Nureyev (the Russian ballet dancer) also died
of AIDS. In the 1990s, tennis player, Arthur Ashe died of AIDS
after a transfusion of HIV-infected blood. Noerine Kaleeba,
Director of the AIDS Support Organisation in Uganda lost her
husband Chris through AIDS; former Zambian President, Kenneth
Kaunda lost his son through AIDS and Fela Kuti, world-renowned
Nigerian musician and political activist died from AIDS. At the
same time, many thousands of nameless people were dying from AIDS
through heterosexual transmission. One name, Nkosi Johnson, became
well known through his brave campaign after he became infected
through mother-to-child transmission. I point out this history to
illustrate that HIV/ AIDS knows no boundaries of class, status,
race or sexual preference. Both the powerful and powerless in every
society are caught up in this vicious epidemic and it is now
estimated that 36 million have been infected worldwide.

A virus named HIV has been identified and fully characterised by
its unique sequence. HIV has fulfilled all of Koch's postulates
[the standard criteria for disease causation] as the sole cause of
AIDS. It is vitally important to recognise that diagnosis and
classification of a disease in medicine is based on the exponential
summation of discriminating characteristics from four components:
medical history, clinical signs, laboratory investigations and
response to treatment. At each level there should be a discriminating
feature that, when taken in context and in toto with the others,
allows us to arrive at a probable diagnosis. This, too, is the case
with AIDS. From this report, it has become clear that statistical
modelling of epidemiological and mortality data adds a fifth
component to the art of diagnosis.

The data presented in this report make the following salient
points:

i) the pattern of mortality from natural causes in South Africa has
shifted from the old to the young over the last decade particularly
for young women - this is a unique phenomenon in biology;

ii) there is a differential mortality pattern between women and
men;

iii) this shift in mortality pattern fits several AIDS models;

iv) the future burden and impact of the epidemic is broadly
predictable from the models with reasonable confidence over the
next decade;

v) the differential patterns of mortality and prevalence will allow
for differential intervention strategies in the different parts of
the country.

This report is a chilling reminder of how powerful stereotypes
across society have colluded in creating the most explosive
epidemic in the history of our country. Comprehensive, powerful and
rigorous as these data are, they can be seized upon positively by
individuals, government and society to intervene at many levels
such that no South African person, family or community has to live
under the cloud of this vicious and unrelenting epidemic.

I sincerely hope that information in this report will be used to
promote the culture of 'Breaking the Silence' around this silent
killer of our nation. As Africa faces the challenges of its renewal
or renaissance, there is no greater potential barrier to the
attainment of this vision than the spectre of the HIV/AIDS
epidemic.

------------------------------------------

Executive Summary

South Africa is experiencing an HIV/AIDS epidemic of shattering
dimensions. The main source of information about the epidemic is
the antenatal clinic HIV seroprevalence surveys conducted by the
Department of Health. Reliable statistics on HIV/ AIDS deaths in
South Africa are not available despite Government's extensive, and
largely successful, efforts to improve the national vital
registration system. The most recent official death statistics
available are those for 1996. By 1996 the proportion of deaths due
to AIDS was too low to tell us much about the shape of things to
come. Even if the numbers of AIDS deaths were substantial, vital
registration statistics may well be an unreliable source of cause
of death information because the true cause of death of someone who
died of AIDS can be expected to be frequently misreported.

Demographic projections of the epidemic indicate that HIV/ AIDS
will cause a rapid change in the age and sex pattern of deaths. A
system to rapidly monitor the age pattern has been developed by the
Medical Research Council. Details of registered deaths are obtained
directly from the Population Register maintained by the Department
of Home Affairs.

Standard indirect techniques have been adapted for estimating the
extent of under-reporting of deaths to allow for different levels
of completeness at different ages which can be expected in South
Africa, in order to estimate the extent of under-registration in
both the routine vital statistics reported by Stats SA as well as
the data obtained from Home Affairs. The coverage of adult death
registration appears to have improved from 54% of deaths occurring
in 1990 being reported to 89% of adult deaths (in those older than
15 years) occurring in the 12- month period to the end of June 2000
being reported. This is a clear sign of the success of the
extensive efforts on the part of Government to improve vital
registration. While this system provides good information on
adults, deaths among children are under-represented as a relatively
high proportion of children are not recorded on the Population
Register.

The data show that there has been a steady increase in adult
mortality during the 1990s. The mortality of young, adult women has
increased rapidly in the last few years with the mortality rate in
the 25-29 year age range in 1999/ 2000 being some 3.5 times higher
than in 1985 (see graph [in PDF version on-line]). The mortality of
young men has also increased, however, the pattern suggested that
this may be a combination of a rise during the early 1990s in
injury-related deaths, that typically occur among men in their
twenties, that began to fall in the late 1990s, and a more recent
increase in deaths due to AIDS in a slightly older age group.
Mortality in the 30-39 year age range in 1999/2000 was nearly 2
times higher than in 1985 (see graph), but obviously this is off a
much higher base.

The pattern in the empirical data is largely consistent with that
predicted by models of the AIDS epidemic, in particular the ASSA600
model developed by the Actuarial Society of South Africa,
suggesting that it is reasonable to interpret an increase in young,
adult mortality as being essentially a consequence of HIV/ AIDS. We
looked at alternative explanations for these patterns and found
none of them plausible. In addition, we cite evidence from a number
of sources in support of our interpretation.

While there is inevitably some degree of uncertainty because of the
assumptions underlying both the model and the interpretation of the
empirical data, we estimate that about 40% of the adult deaths aged
15- 49 that occurred in the year 2000 were due to HIV/ AIDS and
that about 20% of all adult deaths in that year were due to AIDS.
When this is combined with the excess deaths in childhood, it is
estimated that AIDS accounted for about 25% of all deaths in the
year 2000 and has become the single biggest cause of death. The
projections show that, without treatment to prevent AIDS, the
number of AIDS deaths can be expected to grow, within the next 10
years, to more than double the number of deaths due to all other
causes, resulting in 5 to 7 million cumulative AIDS deaths in South
Africa by 2010.

This study has demonstrated the value of supplementing the routine
vital statistics with rapid mortality surveillance, making use of
administrative data from the Population Register. The system needs
to be formalized as rapidly as possible with the data being
released routinely to inform research and policy. Further work to
improve models and data is needed to develop the surveillance tool
to meet the needs of provinces and local government and for
assessment of the impact of interventions. Although there is an
impressive consistency between the pattern of total deaths by age
projected by the ASSA600 model and those captured on the Population
Register, the discrepancies suggest that the model can be improved
in a number of ways. Among these it is suggested that no allowance
be made for a reduction in adult mortality since 1985 when
estimating the non- AIDS mortality. In addition, the results
suggest that the estimates of prevalence based on the early
antenatal clinic survey data probably exaggerated the prevalence in
those years. Various other recommendations are made including
extending this work to the provincial level. It is also important
to develop a mechanism to monitor the impact of the AIDS epidemic
on the mortality of children.

The rapid change in the empirical death rates confirms predictions
of the profound impact of AIDS on mortality. These shocking results
need to galvanise efforts to minimize the devastation of the
epidemic.

----------------------------------------------

Introduction

The HIV/ AIDS epidemic in South Africa continues to grow at a rapid
rate. UNAIDS estimates that in 2000, 19.9 % of adults were
infected, up from 12.9 % two years previously. According to UN
figures with an estimated total of 4.2 million infected (and some
put the figure higher than this), South Africa is said to have more
people living with HIV than any other country.

Reliable empirical data on the epidemic in South Africa are hard to
come by. The main source of information is the series of annual
antenatal seroprevalence surveys conducted by the Department of
Health (DoH). These show that South Africa has experienced a very
rapid spread of HIV during the last decade. In 1990, the first year
of the survey, prevalence was less than 1% and by 2000 its level
was nearly 25%. This yearly survey, covering all the regions of the
country, is conducted on a sample of the routine bloods taken from
pregnant women who attend the public health sector for antenatal
care. The majority of pregnant women (over 80%) make use of public
antenatal care. This group of women makes an ideal sentinel group
for monitoring the epidemic as they have recently had unprotected
sex. In 1998 the protocol for this survey was revised to
standardise procedures and sampling methodology across all the
provinces. It was also changed to allow women to choose whether or
not to be tested, however, it is thought that very few refuse. Not
only does this change make it difficult to compare the figures of
more recent years with earlier years, it also makes the recent data
more difficult to interpret. Despite the observed anomalies in the
provincial level results, this survey provides reasonably
consistent data that form a foundation for surveillance of the
epidemic.

Various projections of the demographic impact of HIV, based on the
antenatal survey results, suggest that the disease will have a
considerable impact on mortality in South Africa. While the
projections differ somewhat, they suggest that between 2000 and
2010, somewhere between four and seven million South Africans will
die from AIDS. This number of AIDS deaths will be considerably
larger than that from any other single cause of death and will be
almost double the number of deaths from all other causes combined
over that period. Model projections of the impact of HIV/AIDS have
an important role to play in providing planning information.
However, their accuracy depends on the many assumptions that are
made in the model. Considering the magnitude of the epidemic, it is
extremely important for South Africa to monitor AIDS mortality so
as to provide reliable information for planning and to be able to
assess the impact of interventions.

Routine mortality statistics are compiled by Statistics South
Africa (Stats SA) from the vital registration system. The
statistics are based on the medical certification of the cause of
death, required by law, at the time of registration of the death
with the Department of Home Affairs (DHA). However, the statistics
are problematic, as death registration is known to have been
incomplete and to suffer from misclassification of cause of death.
After 1994, the Government initiated extensive efforts to improve
death registration and statistics. These involved significant
interdepartmental collaboration, the introduction of a new death
certificate, dissemination of manuals on how to complete the death
certificate and classify the cause of death, and the establishment
of a task teams in each province to improve registration.
Comparison with projections from the ASSA600 model (of the
Actuarial Society of South Africa) suggest that the percentage of
all deaths registered improved from a low of slightly more than 50%
in 1990 to 78% in 1995 and over 80% in 1996.

Despite improved registration, delays continue to occur in the
production of full cause of death statistics. The most recent
detailed statistics are for 1996. Furthermore, cause of death
statistics significantly underestimate the number of AIDS deaths.
Due to the stigma associated with HIV and AIDS, details completed
on the certificate tend to focus on opportunistic infections or the
mechanism of death rather than providing the underlying cause.
Thus, the routine official death statistics fail to provide timely
or accurate information on the extent of AIDS deaths and, at best,
give underestimated numbers some time later.

The Department of Home Affairs maintains the national population
register on computer. It comprises administrative details of all
persons who have been issued with a South African identity
document. A unique identity (ID) number is allocated to each
individual on the register. Death details are also included on the
population register. This database provides information on the age
and sex of dead individuals who were on the Population Register.
Since many children are not registered, this data source does not
provide adequate statistics on children.

This study investigates the trends in reported deaths up until 1996
based on the statistics from Stats SA and compares the results for
adults with more recent data obtained from the population register.
After adjusting for the under-reporting of deaths, the trend in the
age pattern and the broad cause of death profiles are considered to
assess the impact of HIV/AIDS on mortality in South Africa. The
empirical data are compared with model estimates based on the
ASSA600 AIDS and demographic model to assess the consistency of the
empirical data with the model projections.

************************************************************
This material is being reposted for wider distribution by
Africa Action (incorporating the Africa Policy Information
Center, The Africa Fund, and the American Committee on Africa).
Africa Action's information services provide accessible
information and analysis in order to promote U.S. and
international policies toward Africa that advance economic,
political and social justice and the full spectrum of human rights.

Documents previously distributed, as well as a wide range of
additional information, are also available on the Web at:
http://www.africapolicy.org

To be added to or dropped from the distribution list write to
apic@igc.org. For more information about reposted material,
please contact directly the source mentioned in the posting.

Africa Action / Africa Policy Information Center (APIC)
110 Maryland Ave. NE, #508, Washington, DC 20002.
Phone: 202-546-7961. Fax: 202-546-1545.
E-mail: apic@igc.org