This article was originally published in the Business Report on 14 September 2006 By Pali Lehohla 

Last week's release of Statistics SA's report on adult mortality was accompanied by extensive discussion and comment.

In some ways, this was surprising, because the data on which the report was based had been in the public domain for some time, and had already been subject to analysis.

On the other hand, the health of citizens is a concern in all countries, and understanding the causes of death is crucial for effective policy planning and intervention to improve rates of survival. Analysis of mortality trends underlies the development of programmes to reduce mortality from HIV/Aids, malaria, tuberculosis, non-communicable diseases such as diabetes, asthma and hypertension, as well as unnatural causes.

In June, Stats SA released data on registered deaths between 1997 and 2004 by age, sex and year of death. Last week's report was based on a subset of this data, analysing death rates by age, sex, year and cause of death for people between ages 15 and 64.

The methodology used to estimate the percent of deaths that are registered is the same as that used to estimate the mid-year population size and the prevalence of HIV, and is used by both the UN Programme on HIV/Aids (UNAids) and the department of health. This methodology applies consistent procedures to produce mid-year population estimates, to estimate the percentage of deaths registered, and to estimate the prevalence of HIV by age and sex.

The co-operative effort between Stats SA, the department of health and UNAids, and the application of a consistent methodology across a number of areas, helps ensure that analyses and conclusions based on the data are the result of real patterns rather than sometimes unsound assumptions about levels and trends in population phenomena.

The latest data on mortality found that death rates rose between 1997 and 2004 for every five-year age group for each sex, except for males between age 15 and 19. Some of the increases in death rates were very large. For example, the death rates more than tripled for females between ages 20 and 39, and more than doubled for males between ages 30 and 44.

For young and old, increases in death rates were smaller. For each sex, for those between ages 15 and 19 and between 55 and 64, death rates between 1997 and 2004 increased by 20 percent or less.

Overall, death rates from infectious diseases more than tripled for males and increased almost fivefold for females between 1997 and 2004. A large part of this increase can be attributed to HIV, where death rates have a distinctive age pattern in which there is an increase to a given age and then a rapid decline at older ages. This peak occurs at between ages 30 and 34 for females, and between 35 and 39 for males.

Many HIV deaths are registered as due to some other cause. Based on the age pattern of death rates by sex, it is likely that a high proportion of deaths registered as due to parasitic diseases and maternal conditions are caused by HIV.

This non-reporting of HIV deaths probably conceals how many actual deaths result from HIV. A challenge for researchers is to find a way to determine how many deaths are actually caused by HIV. Until that has been accomplished, it is not possible to know how many HIV deaths there are by age, sex and year of death.

Between 1997 and 2004 there was virtually no change in overall cancer death rates for either sex, with cancer death rates actually declining for those between 50 and 64.

There has been concern about mortality from malaria, especially as drug-resistant strains of the disease have become common. Malaria death rates rose between 1997 and 1999, and declined between 1999 and 2004, although the overall death rate from malaria in 2004 was 45 percent higher than the 1997 value for males, and the overall death rate from malaria in 2004 was 93 percent higher than the 1997 rate for females.

Diabetes has also been of growing concern. Adoption of a Western diet, with lots of fat and few fruits and vegetables often leads to obesity. This increases the risk of Type 2 diabetes, which can be fatal. For each sex, death rates from diabetes and obesity rose between 1997 and 2004. Female death rates were higher than male death rates, but the gap narrowed over time. Obesity and diet are an increasing concern for all South Africans.

Mortality rates from unnatural causes, such as homicide, changed little between 1997 and 2004, although they declined between ages 15 and 19, and older.

The risk of unnatural death has variations across the seasons. It is 25 percent higher on a day in December than in other months. Thus, the holiday period is especially risky for several unnatural causes of death, including use of firearms, knives, transport, drowning and falls.

The findings in the adult mortality report give more detail of the causes of adult mortality than has been available previously. However, speculation that this reflects a change in the way mortality data are being interpreted is erroneous. On the contrary, it is the consistency of the methodologies used - not only for data, but also mid-year population estimates and HIV prevalence rates - that allow for a more comprehensive profile of our population.

Pali Lehohla is South Africa's statistician-general and head of Stats SA. For more information on Stats SA and its statistical output, including the latest LFS release, visit www.statssa.gov.za, or contact user services on (012) 310-8600