Introduction Interpersonal violence is an epidemic in South Africa and remains an under-reported and expensive burden on health resources. In most of the developing world there is little or no descriptive information about the expense of treating the consequences of interpersonal violence.
Objective To review the direct burden of interpersonal violence on a tertiary hospital in Northern KwaZulu-Natal, an area known to have high rates of poverty and violent crime.
Material and methods A retrospective case note review of emergency hospital admissions between January and March 2013 was carried out. The reports included demographic characteristics, admitting diagnosis and surgical management. Case files were reviewed to determine cost drivers, such as radiological investigations, blood products, theatre usage and specialist care.
Results Trauma accounted for 374 hospital admissions from the emergency department, of which 142 (38%) were attributable to interpersonal violence (16% of total admissions). One hundred and fifty-six hospital bed days were used over the study period. The average inpatient stay was 9.8 days with 58% requiring a resuscitation bed on admission. One-third of patients underwent emergency surgery and eight patients required postoperative intensive care. The minimum hospital expenditure for interpersonal violence injuries over 3 months was R8 367 788 ($783 960).
Discussion Interpersonal violence is the source of a significant financial burden on the South African health system. Patients are often severely injured and require a high level of specialist investigations and surgical care. This study gives evidence to improve budget and workload planning for regional surgical departments and supports the need for more effective primary prevention.
- violence, interpersonal
- emergency care systems, remote and rural medicine
- cost effectiveness
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
- violence, interpersonal
- emergency care systems, remote and rural medicine
- cost effectiveness
What is already known on this subject?
South Africa has one of the highest rates of violent crime worldwide; however, little research on the healthcare burden of this type of trauma has been carried out.
Injuries from interpersonal violence are an under-reported and expensive burden on health and social resources.
Economic data are essential to evaluate the cost-effectiveness of preventative measures, yet few data on costs exist in many developing countries.
What might this study add?
Interpersonal violence accounted for one-third of trauma admissions, with one-third of these requiring emergency surgery.
The average measurable hospital expenditure was US$5528 per patient, a preventable cost on a limited South African healthcare system.
Interpersonal violence is an epidemic in South Africa. It causes an estimated 27 563 deaths a year; seven times the global rate, placing South Africa among the most violent countries in the world.1 These injuries are considered an under-reported and expensive burden on health and social resources2 and in most of the developing world, there is little or no descriptive information about the expense of treating the consequences of interpersonal violence. Many international committees, such as the WHO, have called for more research into the healthcare burden of interpersonal violence and the direct cost of treating these patients.
In this observational study, we review the direct impact of interpersonal violence on a tertiary hospital in Northern KwaZulu-Natal, an area known to have high rates of poverty and violent crime.3–6 The hospital has 114 allocated surgical beds and a small surgical team but is the referral centre for 20 district hospitals and covers an estimated projected population of 2.34 million people.7 In this setting of limited resources, we aimed to quantify the use of public funds and reserves such as blood products and inpatient bed stays as well as the workload burden that interpersonal violence has on a surgical department. Using recent costing data from neighbouring healthcare districts, we estimate the most minimal expense of this preventable ‘disease’ on the South African health system.
Material and methods
All emergency hospital admission reports from the surgical department were reviewed retrospectively between 1 January 2013 and 31 March 2013. The reports were made by the medical officer on call and included all surgical reviews from the 24 on-call period. Patients discharged from the emergency department after surgical review were excluded. The two primary investigators (SB and ID) independently examined all the reports for patient demographic characteristics, admitting diagnosis and details of acute surgical management.
Data were electronically recorded on a password-protected database of chart abstraction forms. Surgical admissions due to interpersonal violence were identified as patients being admitted under the surgical team secondary to purposeful violence generated from another person or group of people. This was first identified by the admission report from the medical officer on call and confirmed after examining the patient notes. Abstraction forms were cross-evaluated by both investigators to ensure the mechanism of injury involved interpersonal violence and to ensure there was complete agreement about the admitting details. Investigators were aware of the study hypothesis before data collection. All identifiable patient data were excluded and patients were categorised according to the mechanism of injury: stab, gunshot wound, burn and blunt injury. Blunt injury was defined as non-penetrating trauma. For the purpose of costing, blunt trauma was further divided anatomically into head and body injury, as the management of these differed greatly.
The term ‘community assault’ was used when patients were admitted to hospital having been assaulted by members of their village or community owing to an alleged crime. This was first identified by the admitting medical officer's report, and injury details were confirmed by checking the patient's notes. These patients were identified in an additional section of the data abstraction form.
Patients transferred to another hospital after resuscitation or readmitted for step-down care were excluded as this did not directly affect the hospital's resources in an acute setting and resource costs often stretched outside the study period. Patients managed by the emergency department without hospital admission or by the outpatient surgical team (who are often advisers to other hospitals) and any person who died before admission were also excluded. Self-inflicted violence, drowning and parasuicide were categorised separately as ‘other trauma’, and patients experiencing solely sexual assault injuries were also excluded from the study as these patients were managed at a separate facility.
Case files were examined for further information, such as details of inpatient stay, need for surgical intervention and the use of hospital resources. The use of radiological investigations, blood products and specialist care was recorded and considered as other cost drivers. The cost of radiological procedures and blood products was taken from the South African Board of Healthcare Funders’ National Health reference price list, and we used hospital pricing lists as well as cost analysis research from neighbouring hospitals,8 including the price of theatre overheads.9
This was a retrospective review of surgical admissions that did not involve patient participation. Case details had already been collected by the department as part of routine case load auditing, but ethical permission to analyse the data was sought and approved by the hospital chief executive officer.
There were 911 acute surgical admissions over 3 months, of which 41% (n=374) were attributable to trauma (figure 1). Interpersonal violence accounted for more than one-third of the trauma load (figure 2) and 16% of total admissions (n=142). The mean age of patients who were victims of interpersonal violence was 28 (SD ±9, range 4–81) and patients were predominately male. Table 1 shows the baseline characteristics and other key findings. The majority of patients were listed as unemployed (78%) and 6% had had previous hospital admissions secondary to violent injury. Incomplete files with ambiguous patient demographics were excluded from the study. We observed a higher incidence of admissions at the weekend and during public holidays.
Stab injury was the leading mechanism of interpersonal violence injury with most wounds being to the chest and 121 (85%) patients required an emergency chest drain for a pneumothorax or haemothorax. Figure 3 shows the breakdown of types of interpersonal violence injuries. Gunshot wounds were the second most common mechanism of violent injury, frequently with multiple injuries, with 15 patients shot in the abdomen (63%). Two patients were admitted with accidental gunshot wounds and these were categorised as ‘accidental injury.’ Blunt injury to the body secondary to community assault accounted for 14% (n=20) of interpersonal violence admissions, of which 14 patients were diagnosed with acute kidney injury accordingly to RIFLE criteria.10 Of these, five patients required renal dialysis in the acute setting. Other injuries included human bites, crush fractures and lacerations.
The average inpatient stay for interpersonal violence patients was 9.8 days (range 1–66) with 58% requiring a resuscitation bed on admission. Community assault patients required resuscitation beds for longer periods (48 h on average, range 12–168 h) than other patients who were victims of interpersonal violence (average 6 h, range 0–72 h). One-third of patients (n=48) underwent emergency surgery (table 2) and of these, eight patients required postoperative intensive care.
Inpatient stay was longest in patients with burns injuries and gunshot wounds to the abdomen. In total, we calculated that 1356 hospital bed days were used over 3 months by these patients (table 3).
Radiological investigations were required for all patients and 50% required blood products. Resource use is described in table 4.
The inpatient mortality rate among patients admitted owing to interpersonal violence was 2%. These were patients who had been admitted with gunshot wounds to the abdomen (n=2) and owing to community assault (n=1).
The total cost of the resources measured for the 3-month study period was R8 367 788—about US$3.14 million annually.
Interpersonal violence is an epidemic in KwaZulu-Natal but the strain on our limited resources is difficult to quantify. Many studies have examined exposure and epidemiological figures but this study demonstrates the direct cost to the healthcare system and that it accounts for a significant proportion of surgical admissions and absorbs a large share of public resources. Most patients were of working age, and represent a potential loss to the work force, thus highlighting the socioeconomic consequences of these injuries, which are thought to impose a much greater cost burden than the direct cost of treating the injury.2
Community assault occurring as a result of a community-determined justice system for an alleged crime is a well-known term among South African doctors and accounted for a significant proportion (14%) of interpersonal violence admissions. During the study period this was secondary only to blunt trauma. However, other mechanisms of injury, such as flame burns and drowning, have also been reasons for surgical admission after community assault. The concept of ‘community assault’ or ‘community justice’ occurs commonly in South Africa and is likely to stem from poor access to law enforcement agencies. For many rural African communities, a history of exclusion from the government justice system during colonial rule will have reinforced community confidence in traditional chiefs, lineages and other social networks.11 Unfortunately, in the rural catchment area of this tertiary hospital, law enforcement is challenging and many households are dissatisfied with the level of policing in their community.12 Some studies indicate that community justice and vigilantism (victims and their supporters engaging in actions to hold perceived offenders accountable) provide an effective response by holding offenders answerable and deterring crime in the poorer communities.13 However, the punishment (justice) is usually in the form of violence, such as beating, kicking or whipping with a ‘sjambok’; a long thin ‘whip’-like instrument made out of rhino hide or leather. It is an outdated system and the level of violence is difficult to control.
‘Community justice’ as an alternative to the conventional justice system is thought to apply in up to 6% of all crimes in South Africa,14 highlighting the extent of the problem and the resultant hospital admissions. Its practice often results in severe soft tissue and muscular trauma, causing rhabdomyolsis and subsequent acute kidney injury.10 Even though the intent is to inflict only injury, deaths, like the one described in this study, are not uncommon. This subgroup of patients required longer time in the resuscitation unit, with five patients requiring renal dialysis demonstrating the severity of the assault. With better policing and stricter law enforcement deadlines there is scope for improvement in this area. However, until communities feel more supported by the justice system, medical practitioners must recognise the significant morbidity associated with this practice.
Injuries due to interpersonal violence were often severe and required a high level of investigation and surgical care. This study shows that many patients admitted owing to interpersonal violence require radiological investigations and blood products, imposing a significant cost on the hospital. The estimated blood product usage in victims of interpersonal violence was 304 units a year; however this is likely to be underestimated as many paper blood transfusion records were lost from files. Blood products are scarce in KwaZulu-Natal with announced shortages occurring in this tertiary centre 2–4 times a month between January and March 2013. Owing to the emergency nature of the request, blood products are often prioritised for surgical patients and so there is a potential impact on other inpatients, although this is difficult to measure. The high use of radiological investigations and blood products also exposes patients to other health risks in addition to possible future complications.15–17
A significant proportion of patients required emergency surgery. During normal working hours this would delay elective surgical lists or result in cancellations. Explorative laparotomy was the most commonly performed emergency operation and was directly associated with a longer hospital stay. Although the average inpatient stay was 9.8 days, this represented stay only at the tertiary centre and not time spent by those patients transferred for higher surgical care in specialist centres (eg, neurosurgical or cardiothoracic intervention) or those returning to their base hospital for nursing care. Only the resource use for acute admissions was recorded and not that of patients requiring subsequent surgical intervention or continuing follow-up by allied health professionals or in a surgical outpatient clinic. Of the total inpatient bed days for interpersonal violence admissions, 11% were in a high-care ‘resuscitation’ bed and 10% was spent in an intensive care bed, which increased hospital costs substantially and directly affected other hospital patients by blocking already limited beds. This often resulted in severely unwell patients being managed on the general ward without immediate specialist support.
It is important to appreciate that the patients in this study are only those admitted acutely by the surgical team and not those managed by casualty, district hospitals or the sexual assault services. Evidence for the burden of disability and chronic morbidity after violent injuries indicates that the costs are much higher than those described in this study.18 ,19 Owing to the functional limitations of the hospital, it was not possible to carry out a detailed follow-up to quantify continuing morbidity and continuing cost to the health system.
Although this study estimated an average cost of R58 928 ($5524) per patient, it was not possible to measure many other cost drivers—for example, staffing costs, operating theatre equipment, transport, medication, specialist nutrition, dressings and hospital utility bills. Additionally, this was a retrospective study based on the accuracy of handwritten charts and it is possible that patients were accidentally excluded either through human error or misdiagnosis. A more accurate cost analysis would involve microcosting with independent abstractors, which is considered to be the ‘gold standard’ for costing inpatient stays.20 Neighbouring hospitals have shown that ward overheads were the most expensive aspect of inpatient stay8 and this was not investigated, contributing to the underestimation of expenditure. During the study period, 17 patient notes could not be found to confirm diagnosis and so were excluded from the study. These limitations suggest that the cost of interpersonal violence is much higher and a prospective review will be a useful follow-on study.
Interpersonal violence is a public health problem in KwaZulu-Natal and the solution to this significant burden lies within the rural social structures of South Africa. Primary prevention is possible and trials of small-scale interventions, such as community education, gun safety training and home visitation services, have been carried out in some communities,21 but lack robust evidence for effectiveness. These strategies require greater investment with intervention required at all levels of the ecological model, including large-scale involvement of government, community heads and police, in order to achieve an effective outcome.
Contributors SB and MN designed the study. SB and ID collected data and were responsible for its interpretation. SB drafted the article, which was further edited by ID and CA. All authors reviewed the manuscript and approved the final version for publication.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement At present, the study data are unpublished and only available to the article authors. However, all authors agree that this is an important subject for which data should be available to encourage ongoing work and we would be willing to make all study data available.