FACT CHECK: Inside the figures for gun-related deaths in South Africa
By Tarryn Jane Roy
South Africans have in recent months debated changes proposed by the Police Ministry to the country’s main firearms law.
Among the most controversial of the changes is the removal of self-defence as a valid reason for owning a firearm.
According to 2019/20 police statistics, firearms are the most commonly used murder weapon. But what else do we know about gun deaths in South Africa?
This factsheet takes a closer look.
WHAT ARE GUN DEATHS?
A “gun-related death” is any death from a gunshot wound, Professor Daniel Webster, a gun policy expert at Johns Hopkins University in the US, told Africa Check.
He told Africa Check that there were two main categories of gun death. These are:
Intentional (murder and suicide)
But he added that “undetermined intent” is another important category to consider because intent is not always clear when the death certificate is issued.
HOW ARE THESE COUNTED?
Public health researchers use death certificate data to sort between murders, suicides, legal interventions, unintentional deaths and gun deaths of undetermined intent, Webster said.
Matzopoulos said that in South Africa, death certificate data and post-mortem reports signed off by forensic pathologists and doctors are the two key sources of data on gun deaths.
South African Police Service (SAPS) crime statistics now indicated the murder weapon, he said, so rates of firearm homicide could be calculated.
PROBLEMS WITH GUN DEATH DATA
Misclassification and under-reporting could be an issue in many sources of data, Matzopoulos said. Misclassification is incorrectly assigning a gun death to one of the categories of intentional, unintentional, or undetermined.
“In South Africa non-natural deaths are subject to huge misclassification issues,” he said.
A death certificate will indicate whether the death was natural or not. If it is unnatural, the death certificate will specify the external cause of death and whether the death was intentional or unintentional or of undetermined intent.
But “intentionality is difficult for them to figure out at the time of death and so they often describe the intent as undetermined”, Matzopoulos said.
A June 2017 paper from the burden of disease unit published in the South African Medical Journal says that this is a complication because, according to international coding norms, a death must be classified as unintentional when the intention is unknown.
So, when death certificate data is used, intentional gun deaths are often underreported. By examining data from 1997 to 2013, Matzopoulos and his co- researchers found wide misclassification of gun deaths.
“That’s why we can't use it at face value. We have to do studies to address the misclassification,” Matzopoulos said.
Credible data would adjust for underreporting and misclassification, he added.
SOURCES OF DATA ON GUN DEATHS
South African Police Service
The SAPS releases crime statistics for each quarter and each year.
Statistics South Africa
Stats SA publishes annual data on mortality and causes of death in South Africa. This, it says, is based on deaths collected through the country’s civil registration system maintained by the Department of Home Affairs.
The statistics agency says cause of death information “is as recorded on death notification forms completed by medical practitioners and other certifying officials”.
If the intention is unclear, they are recorded in the external causes of accidental injury category, under exposure to inanimate mechanical forces. This subcategory includes other causes of death such as knives and swords.
SAMRC’s national burden of disease study
These estimates are compared to official statistics and those from the global burden of disease study. The latter is maintained by the Institute for Health Metrics and Evaluation or IHME, an independent global health research centre at the University of Washington in the US.
The most recent study was for 2009. Another, for 2017, which would be the second, is on the way, Matzopoulos said.
“Fieldwork is now complete with the main report due to be released in the next couple of months after completing peer review and final edits.”
Gun death data is collected from various sources depending on the country. The school’s total deaths for South Africa are from the World Health Organization’s mortality database. But this link is no longer accessible.
World Health Organization
The WHO maintains a database on mortality. But Stephanie Burrows, technical officer at the WHO violence prevention unit, told Africa Check that data on firearm-specific deaths is not easily available from this database as such information is not routinely extracted.
The agency did have “basic underlying data” for 2015 that could be downloaded, but specific expertise was needed to use it.
The 2015 data is also complicated by different countries using different international codes to sort firearm deaths.
IHME global burden of disease study
IHME spokesperson Amelia Apfel told Africa Check that the data covers three gun death categories: interpersonal violence, self-harm by firearm and unintentional firearm injuries.
To make these estimates, the institute uses statistical modelling. This means they don’t “provide data in the true sense” but rather “provide estimates derived from different data points”, explained Matzopoulos.
Stats SA’s vital registration data is a key data point for the model, meaning its estimates are often lower than official figures such as crime statistics.
The SAMRC has in the past informed the study, “resulting in quite substantial changes to the earlier iterations of their model”, Matzopolous said.
But the IHME “do refine their models all the time” and “one would hope that the model is making better adjustments for this misclassification as time progresses”.
United Nations Office of Drugs and Crime
The small arms survey
Andrei Muchnik, a communications officer from the WHO, referred us to this survey as a source of data. The small arms survey is an annual project by the Graduate Institute of International and Development Studies in Geneva, Switzerland.
It keeps an annual global violent deaths database.
Nicolas Florquin, the head of the data and analytics unit at the survey, told Africa Check that the last year their most recent data came from was 2018.
WHAT DO THE NUMBERS SHOW?
Epidemiologists also take into account different age structures, because these can have an impact on the data. For example, an older population will have a higher overall mortality rate than a younger population. So, when comparing data, age would be converted to a uniform standard.
The most recent annual police crime statistics are for 2019/2020, when 7,351 gun murders were reported. This works out to a rate of 12.5 gun murders per 100,000, using Stats SA population data for the end of September 2019.
The SAMRC’s burden of disease unit gave us their preliminary results for the third national burden of disease injury mortality survey. These were presented at the WHO’s virtual 2021 world injury conference.
The study recorded 6,275 gun homicides. Using Stats SA population estimates, the rate is 11.1 gun murders per 100,000 in 2017.
When compared with the murder rates from other sources, the IHME global burden of disease study’s rates are substantially lower.
Megan Prinsloo is a specialist scientist for the SAMRC’s burden of disease research unit. She told Africa Check that according to the 2017 injury mortality survey, guns account for “approximately 7%” of suicide deaths.
Preliminary data from the study counted 453 gun suicides. This works out to a rate of 0.8 deaths per 100,0000 using Stats SA population estimates.
UNINTENTIONAL AND UNDETERMINED GUN DEATHS
The most recent unintentional deaths figure available for South Africa is from the global burden of disease study for 2018. This estimates 276 unintentional gun deaths, or a rate of 0.5 deaths per 100,000 people.
The 2017 injury mortality survey found that “firearms account for less than 1% of other unintentional and undetermined deaths”.
Using Stats SA population data, we calculated the unintentional and undetermined gun death rate per 100,000 people to be 0.1 gun deaths per 100,000 people.
Gunpolicy.org recorded a much higher rate of 10.25 unintentional deaths per 100,000 in 2015 and 0.01 gun deaths of undetermined cause.
OVERALL GUN DEATHS FIGURES
The most recent overall gun deaths data for South Africa that we found was from the IHME, which counted 3,610 in 2019.
Although gun death figures based on mortuary records for 2015/16, 2016/17, and 2017/18 were given in parliament, these only included figures from four of South Africa’s nine provinces: Gauteng, Limpopo, the Western Cape and an unnamed fourth province. So we could not use this data.
We collated the available data on gun death rates from various sources into a table.
A figure of 6,258 gun deaths recorded by all state mortuaries in 2014 was given to parliament. Using Stats SA’s 2014 population estimates we worked the rate out to be 11.6 gun deaths per 100,000.
The 2017 injury mortality preliminary results showed 6,768 gun deaths. This is 12 deaths per 100,000 people, using Stats SA population estimates.