Levels and factors associated with homicide-related deaths in a rural South African population
Date
2011-03-25
Authors
Otieno, George Omondi
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Background: World Health Organization (WHO) estimates that more than 1.6 million people die every year because of violence and out of these deaths, homicide accounts for almost one third. Ninety percent (90%) of homicide are thought to occur in low and middle income countries. South Africa has one of the most disturbing rates of homicide in the world. These high homicide rates besides resulting in reduced life expectancy also have serious health, social and economic consequences. Aim: The study aimed at quantifying the burden as well as and identifying factors associated with homicide deaths in rural KwaZulu-Natal in South Africa during the period of 2000 to 2008. Objectives: To estimate a 9 year period (2000-2008) homicide incidence rates as well as identify factors associated with homicide-related deaths. Further, the analysis described spatial distribution of homicide-related deaths in a rural South African population. Design: Analytical longitudinal study. Methods: Using data drawn from the Verbal Autopsies (VAs) conducted on all deaths recorded during annual demographic and health surveillance over a 9-year period (2000-2008), Kaplan-Meier (K-M) survival estimates of incidence rates were used to estimate the cumulative probability of death until the end of the period. Estimates were reported by sex and residency. Weibull regression methods were used to investigate factor associated with homicide deaths. Kulldorff spatial scan statistics was used to describe homicide clustering.
Results: With 536 homicide-related deaths, and 814, 715 total Person Years of contribution, the study found an overall incidence rate of 66 (95% CI= (60, 72) per 100, 000 Person Years of
v
Observation (PYOs) for the period studied. Death due to firearm was reported the leading cause of mortality (65%). Most deaths occurred over the weekends (43%), followed by Friday (16.2%).The highest homicide incidence rates were recorded in 2001 (90; 95% CI= (71, 111) per 100,000 person years at risk and 2004 (86; 95% CI= (68, 108) per 100,000 person years at risk. Males had a rate that was about six times more than females 115 (95% CI=105,127) per 100,000 PYOs. Age-specific homicide rate were highest among males aged 25-29 years (209.90 per 100,000 PYOs) and females aged 50-54 years (78 per 100,000 PYOs). Resident, age, sex, education, socioeconomic status, and employment independently predicted homicide risk. The study identified two geographical clusters with significantly elevated homicide risk. Conclusion: A significant six fold difference in homicide rate existed between males and females. Sex differential increases with age, with males aged 15-54 years the most likely to be killed, and females aged 55 years and above having the highest homicide rate. Increase in wealth status and level of education increases one‘s risk of homicide. Employment per se was protective from homicide risk. Firearm was the leading cause of mortality. Most deaths occur over the weekend. Two geographical areas with elevated homicide risk were observed. These findings underscore the need to have timely information and strategies for effective violence prevention program to subgroups and areas at risk.
Description
MSc (Med), Population-Based Field Epidemiology, School of Public Health, Faculty of Health Sciences, University of the Witwatersrand
Keywords
homicide, rural, contributing factors