SciELO - Scientific Electronic Library Online

vol.100 issue12 author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand



Related links

  • On index processCited by Google
  • On index processSimilars in Google


SAMJ: South African Medical Journal

On-line version ISSN 2078-5135
Print version ISSN 0256-9574

SAMJ, S. Afr. med. j. vol.100 n.12 Pretoria Dec. 2010




Canada's health care system: a relevant approach for South Africa?



Anne-Emanuelle BirnI; Stephanie NixonII

IMA, ScD. Dalla Lana School of Public Health, University of Toronto, Canada
IIPhD. Department of Physical Therapy, University of Toronto, and Health Economics and HIV/AIDS Research Division (HEARD), University of KwaZulu-Natal, Durban




BACKGROUND: While countries such as the USA, South Africa and China debate health reforms to improve access to care while rationalising costs, Canada's health care system has emerged as a notable option. In the USA, meaningful discussion of the advantages and disadvantages of the Canadian system has been thwarted by ideological mudslinging on the part of large insurance companies seeking to preserve their ultra-profitable turf and backed by conservative political forces stirring up old fears of 'socialised medicine'. These distractions have relegated the possibility of a 'public option' to the legislative dustbin, leaving tens of millions of people to face uninsurance, under-insurance, bankruptcy and unnecessary death and suffering, even after passage of the Obama health plan. While South Africa appears to experience similar legislative paralysis, there remains room for reasoned health reform debate to address issues of equity, access, and financing.
OBJECTIVE:. Our aim is to contribute to the debate from a Canadian perspective, setting out the basic principles of Medicare (Canada's health care system), reviewing its advantages and challenges, clarifying misunderstandings, and exploring its relevance to South Africa. We periodically refer to the USA because of the similarities to the South African situation, including its health care system, which mirrors South Africa's current position if left unchanged.
CONCLUSION:. While Medicare is neither flawless nor a model worthy of wholesale imitation, we contend that open discussion of Canada's experience is a useful component in South Africa's current policy and political efforts.



'Of all the forms of inequality, injustice in health care is the most shocking and inhumane.'
  Martin Luther King, Jr

As various large countries (e.g. the USA, South Africa, China) debate health reforms about improved access to care while rationalising expenditures, Canada's health care system has emerged as a prominent option. In the USA, meaningful discussion on the advantages and disadvantages of the Canadian system has been thwarted by ideological mudslinging by defenders of large insurance companies seeking to preserve their ultra-profitable turf and backed by conservative political forces who have stirred up old fears of 'socialised medicine'. These distractions have relegated even the possibility of a 'public option' to the legislative dustbin, leaving tens of millions to face uninsurance, under-insurance, bankruptcy, and unnecessary death and suffering, even after passage of the Obama health plan.

While South Africa appears to be undergoing similar legislative paralysis,1 the opportunity still remains for reasoned health reform debate to address issues of equity, access, and financing. Our aim is to contribute to the debate from a Canadian perspective. We set out the basic principles of Medicare (Canada's health care system), review its advantages and ongoing challenges, clarify misunderstandings, and explore its relevance to South Africa. We periodically refer to the USA because of the similarities to the South African situation, including its health care system which mirrors South Africa's current position, if left unchanged (Table I). While Medicare is neither flawless nor a model capable of wholesale imitation, we contend that open discussion of Canada's experience is a useful component in South Africa's current policy and political efforts.


Founding and principles of the Canadian health care system

Until the 1940s, health care in Canada was primarily delivered through fee-for-service (i.e. direct payment to the service provider) private medical practice. Canada has a decentralised political system with 10 provinces and 3 territories. Regional efforts, starting with the province of Saskatchewan, began a 25-year struggle for national health insurance. Saskatchewan passed a public hospital insurance plan in 1947. After the federal government offered in 1957 to share the costs of providing hospital and diagnostic services, the rest of the country followed suit within 5 years. However, the real struggle proved to be public insurance for physicians' services, passed by the Saskatchewan legislature in 1962. Backed by the North American medical establishment, the province's doctors went on a 23-day strike to forestall the legislation, but the province prevailed and, by 1968, a national Medical Care Act was in place, providing federal funding to all provinces that passed universal hospital and medical care legislation. By 1972, Canada's universal health care system, named Medicare, was fully in place.

While this universal public health insurance plan was highly popular, serious concerns around equity surfaced within a few years because doctors and hospitals charged extra user fees for faster or special services. A review panel concluded that the extra charges were leading to a de facto 2-tiered system that was limiting access to care. In 1984, parliament passed the Canada Health Act (one of the last actions of Prime Minister Pierre Trudeau's administration), setting out 5 principles underpinning Medicare (Table II).

Using federal funding as both carrot and stick, the Canada Health Act sought to ensure equity by penalising extra charges and creating an interlocking system of provincial and territorial health plans, even as financing and coverage vary (see below).

How universal health insurance in Canada works

Canada operates under a single-payer health care system. In contrast to South Africa or the USA, where both private insurance and government schemes co-exist, covered health care services in Canada are paid for through a single source: the provincial/territorial government. Table III summarises single-payer v. multiple-payer health systems.

In each province/territory, the Ministry of Health approves and funds a global annual operating budget for each hospital based on past budgets, patient load, inflation and salary increases, case mix, capital expenditure needs, and other factors. The global budget helps to control costs, as hospital administrators and department directors know they need to stay within the prescribed budget. When unforeseen events occur (e.g. severe acute respiratory syndrome (SARS) in 2003), the province reimburses hospitals that have incurred extra expenditures.

Physicians in private practice (the vast majority in Canada) are paid on a fee-for-service basis with a ceiling on earnings.

Fees are determined through negotiations between provincial medical associations and the corresponding Ministries of Health. Nurses' salaries are negotiated through collective bargaining. Hospital-based physicians and other health workers are paid salaries.

All legal residents of Canada are eligible for health insurance coverage after a minimal waiting period - usually a few months. In accordance with the Canada Health Act, residents with a health care card have access to all covered services with no extra payments or user fees. People in Canada can select their own doctors - a far cry from private insurance plans in the USA, which limit access to doctors according to an ever-changing list.


Two-thirds of Medicare expenditure is funded by the province, based on a variety of mechanisms. General revenues (i.e. from income and corporate taxes) are an important source of revenue. Three provinces (Alberta, Ontario and British Columbia) use premiums based on a sliding income scale. Even those who have not paid the premium, however, are not denied access to health services. Other provinces raise revenues through lotteries and 'sin taxes' on alcohol and cigarettes.

Since 2004, the federal portion of funding has come in a dedicated block grant called the Canada Health Transfer, which the provinces must use for health care expenditure. On average, the transfer funds a third of health care expenditure. To take into account the fiscal disparities among provinces, the federal government also provides equalisation payments to those provinces whose fiscal capacity is below the average of all 10 provinces. In 2009/2010, over $Cdn 14 billion in equalisation transfers is being distributed to 6 provinces for health and other social spending. In addition, the three northern territories receive separate federal transfers ($Cdn 2.5 billion in 2009/2010) for social services for their residents at a comparable level and tax rate as the provinces. Social service delivery in these regions is a considerable challenge because of the large number of small and isolated (snowbound for much of the year) communities.

In 2008, a total of $Cdn 172 billion was spent on health care in Canada, which is approximately $5 170 per capita. While among the highest levels in the world, partly owing to relatively high physicians' salaries, the 10% average of Canada's GDP that goes to health care is roughly one-third less than that of the USA. Furthermore, according to the Organisation for Economic Co-operation and Development (OECD), per capita spending on health care in Canada is 46% lower than in the USA.6 Almost a third of the higher spending in the USA is due to the multiple payer system bureaucracy (as in South Africa currently). Hospital and physician office billing administrative costs are more than three times higher in the USA than in Canada.7

The cost control advantages of Canada's single payer system are evident in Fig. 1. The USA and Canada had virtually the same pattern of health expenditure as a percentage of GDP until the early 1970s. But after Canada's national health insurance programme was fully implemented, it experienced far lower increases in health spending as a percentage of GDP than in the USA.



Ongoing challanges

Limits to coverage

In total, 70% of health care expenditure is publicly funded in Canada. Payers other than government (i.e. private insurance, employers and individuals) fund the remaining 30%. This is because, notwithstanding the principles of universal coverage and comprehensiveness under the Canada Health Act, not all medically necessary services are covered. While the omissions vary by province/territory, services typically left out include eye care (other than for minors and seniors), pharmaceuticals (except for seniors and people with disabilities), long-term care and dental care. These non-covered services have driven most of the rises in health expenditure as a percentage of GDP, with covered hospital, physician and administration costs remaining stable since Medicare was fully implemented.

Brain drain

Canada also experiences a shortage of medical personnel as does South Africa. The shortage of primary care personnel is particularly acute in many rural areas and among Canada's Aboriginal populations. Ironically, in Saskatchewan, which pioneered Medicare, 54% of doctors received their medical training overseas, and almost 20% of doctors were trained in South Africa alone.8 In that sense, there is a direct link between the future sustainability (and equity) of both Canada's and South Africa's health care systems.

While the shortage (or imbalanced distribution) of medical personnel in Canada undoubtedly indicates poor human resources training and management strategies, it is not a problem of Medicare per se. Rather, fault lies with national and regional educational and training policies, most notably an early 1990s policy reducing medical school enrolments by 10%. As such, the shortage must be addressed at the political level, e.g. by raising the number of medical school entry slots, which in recent years have had more than 5 applicants for every medical school place, with large numbers of qualified candidates rejected. The nursing shortage has been even greater than the physician shortage, also leading to reliance on foreign health personnel. Canada has been in part responsible for the 100 000 nurses who emigrated from the Philippines between 1990 and 2004;9 and some 50% of Ghana's nurses have immigrated to Canada.


In the 1990s, waiting lists for non-emergency and specialist hospital services grew because government funding cuts led to the closing of 30% of hospital beds. The 2002 Commission on the Future of Health Care in Canada (the Romanow Commission)10 highlighted such flaws. Broad political consensus led to the disbursement of billions of federal dollars to the provinces to correct this problem, resulting in a dramatic drop in waiting times despite misleading claims by conservative think-tanks. Most importantly, health system-dependent health outcomes and quality of care are notably better in Canada than in the USA.11


Equity considerations include how fairly a health system's resources are financed and distributed, and what role the health system plays in improving equity in society generally.

Canada's national health insurance system is designed to be equitable in both its financing arrangements and its coverage guarantee. While access and resource availability are problematic among rural and northern populations, the federal government and provinces are attempting to redress this problem. In 2005, the Northern Ontario School of Medicine was founded, specifically to attract students from northern and Aboriginal backgrounds to train in community health. As the country's first school of medicine with a social accountability mandate, it is committed to serving rural and remote populations.

At the same time, the Canadian system's principles of universality are being challenged. A 2005 court case in Quebec opened the door for private insurance and private clinics to cover publicly available services that are not available timeously. Because of the Canada Health Act's prohibition on extra billing if provinces want to receive federal subsidies, 6 provinces (including Quebec) have passed laws that prohibit private insurance for medically necessary services that are covered by provincial health insurance.

So far, the importance of maintaining equitable, universal public health insurance has impeded the creation of a two-tier medical care system. But popular support for Medicare hinges on citizen and provider participation in identifying and addressing weaknesses of the system as well as in government responsiveness to improving the quality of services. The federal investment in reducing waiting times in the wake of the Romanow Commission is an example of how broad participation in health policy-making has helped to strengthen Medicare. Any health system must undergo continuous reform to meet shifting population needs and evidence-based developments in medicine and public health services.

The second aspect of equity concerns the role played by the health system in reducing social inequalities (in health). South Africa is one of the most unequal societies in the world, with the richest tenth of the population receiving over half of all income, and the poorest tenth accounting for just 0.2% of income.12 Inequality extremes in Canada are about half of those in South Africa, yet poverty - especially the proportion of children living in poverty - remains entrenched above 10%. Nevertheless, Canada's social welfare system, including income transfers for the elderly and the health care system, has played a significant role in promoting equity.

Social class and racial/ethnic patterns of health expense-related bankruptcies and deaths linked to lack of coverage in the USA provide evidence for the role of health insurance and social welfare programmes generally in reducing social inequalities in health. Medical debts in 2007 made up over 62% of all bankruptcies, up from 46% just 6 years earlier.13 Lack of health insurance is estimated to result in approximately 18 000 deaths per year in the USA.14 Eliminating such sources of inequality should serve as a key incentive for any reform of the health care system in South Africa.


Relevance of the Canadian experience to South Africa

South Africa faces considerable health problems, including rising child mortality and high rates of HIV/AIDS and tuberculosis. These stem from broader societal determinants of health (including poor housing conditions, low wages, unemployment, systemic poverty and inadequate schooling), that can only be partially addressed by the health care system. Health care is but one determinant of health. Nevertheless, a well-functioning and equitable health care system is crucial for society, and health reform in South Africa has gained symbolic value as a sign of socio-political change.

On the surface, it is easy to identify differences between Canada and South Africa that challenge the relevance of a Medicare model in a South African context. A key difference, for example, is the labour force structure in the 2 countries: the proportion of South Africans employed in the formal sector (and hence South Africa's tax base) is far smaller than in Canada. Yet significant parallels between South Africa and Canada inform reflection on the Canadian system during South Africa's health reform deliberations. Over half of South Africa's total health expenditure is spent in the private sector.15 Canada's health system was also mostly private before Medicare evolved. South Africa's health governance comprises 1 national and 9 provincial departments of health, based on a district model. Canada's health system is also decentralised, with supervision provided at the federal level and health services administered and delivered by the provinces and territories.

Like South Africa today, health service provision and health system-related outcomes in Canada before Medicare varied dramatically across its provinces and territories. This inequity has been largely addressed under Medicare. Canada and South Africa have hard-to-reach populations in rural areas that require special targeting within a universal system. Labour unions are influential in both countries, and it is noteworthy that the national health programme has consistently received impassioned support from Canada's labour representatives, as in South Africa.16 Both countries share an explicit commitment to equity for all citizens, which includes focusing on the poorest and most vulnerable in our societies. To this end, not only has Canada's universal health system provided a vehicle for promoting equity, but the principles of Medicare have also become a central part of the Canadian identity.

Perhaps most compelling is the fact that Canada's universal health care system required 25 years of struggle by advocates within government and progressive political parties, the health sector, patient groups, labour unions, certain sectors of business and, importantly, the general population. Medicare was built by Canadians, and requires the ongoing participation of citizens to ensure its relevance as the needs of the diverse population and discoveries in medicine evolve.

The political change witnessed during the advent of Canada's Medicare, however, is dwarfed by the accomplishments of South Africa in the last two decades. Specifically, since 1994, South Africa has made remarkable gains in terms of health care access, including efforts to redress historical inequities and provide essential care to disadvantaged populations.17 The drive to transform South Africa's health care system into one based on equity and a sustainable financial model has been underway since the ANC's health plan was launched 15 years ago. Current deliberations have the opportunity to build on this history of achievements and contribute not only to the health of residents but also to the well-being of the country. The Canadian approach is not a panacea for South Africa, yet lessons may be found in its fundamental principles and funding structures, which can inform South Africa's quest for a fair, effective and equitable health care system.

In 2004, the South African Broadcasting Corporation (SABC) conducted a nationwide poll to determine 'The Greatest South African' of all time. The winner, unsurprisingly, was Nelson Mandela. Something that current South African President Jacob Zuma might consider is that in the same year, the Canadian Broadcasting Corporation (CBC) conducted a parallel poll in Canada. Tommy Douglas, best known as the father of Canada's national health insurance programme, was voted 'The Greatest Canadian' of all time.18

The authors are grateful to Robert Chernomas and Mark Hunter for their helpful comments. Anne-Emanuelle Birn's work on this article was supported, in part, by the Canada Research Chairs programme, which had no role in the writing of the article or the decision to submit it for publication.



1. Zille enters NHI debate. Health-e 18 June 2009. (accessed 19 October 2009).         [ Links ]

2. York G. Can South Africa copy Canada's medicare? Globe and Mail 11 August 2009.         [ Links ]

3. Davie E. Stepping carefully towards health reform. Health Policy Unit. (accessed 14 February 2010).         [ Links ]

4. Mchangama J. Beware the blind alley of the 'right' to health. Business Day 4 January 2010. (accessed 14 February 2010).         [ Links ]

5. Esmail N, Walker M. High-priced Canadian health care system fails patients. Health Policy Unit. (accessed 14 February 2010).         [ Links ]

6. Organization for Economic Cooperation and Development. OECD Health Data 2009: How does Canada compare. (accessed 19 October 2009).         [ Links ]

7. Woolhandler S, Campbell T, Himmelstein DU. Costs of health care administration in the United States and Canada. N Engl J Med 2003; 349: 768-775.         [ Links ]

8. Labonte R, Packer C, Klassen N, et al. The brain drain of health professionals from sub-Saharan Africa to Canada. In: African Migration and Development Series No. 2 (South African Migration Project-SAMP). Cape Town: Idasa Publishing, 2006.         [ Links ]

9. Joint Learning Initiative. Human Resources for Health: Overcoming the Crisis. Cambridge, MA: Harvard University Press, 2004.         [ Links ]

10. Romanow RJ. Building on Values: The Future of Health Care in Canada. Ottawa: Government of Canada Publications, 2002. (accessed 18 October 2009).         [ Links ]

11. Guyatt GH, Devereaux PJ, Lexchin J, et al. A systematic review of studies comparing health outcomes in Canada and the United States. Open Medicine 2007; 1(1): e27-e36.         [ Links ]

12. Coovadia H, Jewkes R, Barron P, Sanders D, McIntyre D. The health and health system of South Africa: historical roots of current public health challenges. Lancet 2009; published online 25 August 2009. DOI:10.1016/S0140-6736(09)60951-X.         [ Links ]

13. Himmelstein DU, Thorne D, Warren E, Woolhandler S. Medical bankruptcy in the United States, 2007: results of a national study. Am J Med 2009; advance e-print.         [ Links ]

14. Institute of Medicine. Insuring America's Health: Principles and Recommendations. Washington, DC: The National Academies Press, 2004. (accessed 12 April 2010).         [ Links ]

15. Chopra M, Daviaud E, Pattinson B, Fonn S, Lawn JE. Saving the lives of South Africa's mothers, babies, and children: can the health system deliver? Lancet 2009; published online 25 August. DOI:10.1016/S0140-6736(09)61123-5.         [ Links ]

16. General Motors, Ford and Daimler-Chrysler and the Canadian Auto Workers. General Motors, Ford and Daimler-Chrysler and the Canadian Auto Workers Support Canadian Medicare: Joint Letter on Publicly Funded Health Care 12 September 2002; and Congress of South African Trade Unions (COSATU) Support for NHI (2000). (accessed 19 October 2009).         [ Links ]

17. Coovadia H, Jewkes R, Barron P, Sanders D, McIntyre D. The health and health system of South Africa: historical roots of current public health challenges. Lancet 2009; published online 25 August. DOI:10.1016/S0140-6736(09)60951-X.         [ Links ]

18. The Greatest Canadian. (accessed 19 October 2009).         [ Links ]



Accepted 22 February 2010.



Corresponding author: A-E Birn (

^rND^sYork^nG^rND^sDavie^nE^rND^sMchangama^nJ^rND^sEsmail^nN^rND^sWalker^nM^rND^sWoolhandler^nS^rND^sCampbell^nT^rND^sHimmelstein^nDU^rND^sLabonte^nR^rND^sPacker^nC^rND^sKlassen^nN^rND^sGuyatt^nGH^rND^sDevereaux^nPJ^rND^sLexchin^nJ^rND^sCoovadia^nH^rND^sJewkes^nR^rND^sBarron^nP^rND^sSanders^nD^rND^sMcIntyre^nD^rND^sHimmelstein^nDU^rND^sThorne^nD^rND^sWarren^nE^rND^sWoolhandler^nS^rND^sChopra^nM^rND^sDaviaud^nE^rND^sPattinson^nB^rND^sFonn^nS^rND^sLawn^nJE^rND^sCoovadia^nH^rND^sJewkes^nR^rND^sBarron^nP^rND^sSanders^nD^rND^sMcIntyre^nD^rND^1A01^nN^sSiegfried^rND^1A01 A02^nJ^sVolmink^rND^1A03^nA^sDhansay^rND^1A01^nN^sSiegfried^rND^1A01 A02^nJ^sVolmink^rND^1A03^nA^sDhansay^rND^1A01^nN^sSiegfried^rND^1A01 A02^nJ^sVolmink^rND^1A03^nA^sDhansay



Does South Africa need a national clinical trials support unit?



N SiegfriedI; J VolminkII; A DhansayIII

IMB ChB, MPH (Hons), FCPHM (SA), DPhil (Oxon). South African Cochrane Centre and MRC Clinical Trials Initiative, Medical Research Council, Cape Town
IIBSc, MB ChB, MPH, DPhil (Oxon). South African Cochrane Centre and MRC Clinical Trials Initiative, Medical Research Council, and Faculty of Health Sciences, Stellenbosch University
IIIMB ChB, DCH, MMed (Paed), FCPaed. Office of the Vice-President: Research, Medical Research Council, and MRC Clinical Trials Initiative, Medical Research Council




BACKGROUND: No national South African institution provides a coherent suite of support, available skills and training for clinicians wishing to conduct randomised controlled trials (RCTs) in the public sector. We report on a study to assess the need for establishing a national South African Clinical Trials Support Unit.
OBJECTIVES: To determine the need for additional training and support for conduct of RCTs within South African institutions; identify challenges facing institutions conducting RCTs; and provide recommendations for enhancing trial conduct within South African public institutions.
DESIGN: Key informant interviews of senior decision-makers at institutions with a stake in the South African public sector clinical trials research environment.
RESULTS: Trial conduct in South Africa faces many challenges, including lack of dedicated funding, the burden on clinical load, and lengthy approval processes. Strengths include the high burden of disease and the prevalence of treatment-naïve patients. Participants expressed a significant need for a national initiative to support and enhance the conduct of public sector RCTs. Research methods training and statistical support were viewed as key. There was a broad range of views regarding the structure and focus of such an initiative, but there was agreement that the national government should provide specific funding for this purpose.
CONCLUSIONS: Stakeholders generally support the establishment of a national clinical trials support initiative. Consideration must be given to the sustainability of such an initiative, in terms of funding, staffing, expected research outputs and permanence of location.



The randomised controlled trial (RCT) is the gold standard for evaluating the effects of health care interventions.1 Therefore, South African health policy and clinical guidelines should be based on well-conducted RCTs that ideally are conducted within the country to ensure local applicability.

Conducting RCTs in South Africa faces numerous obstacles. Specialist training in the universities traditionally focuses on clinical experience and skills accrual, and lacks a research focus. Downscaled tertiary service units struggle to remain academically active,2 reducing the opportunities for local clinicians to acquire the epidemiological and statistical skills for conducting RCTs. Health professionals interested in clinical research may have to migrate to develop these skills. Opportunities to then practise those skills may only exist overseas, contributing to the professional brain drain.3

When research funding is not available from local sources, researchers become dependent on funds from donor agencies or the drug industry, potentially deflecting priorities away from local needs.4 African governments have committed to spend 2% of their health budgets on research.5 In 2008/2009, the South African Medical Research Council (MRC) received far short of the promised 2% from the national Department of Health (DoH), even if National Research Foundation monies are added to this.6 The MRC is reported to be chronically underfunded despite its mandate to maintain and develop clinical research capacity.2

RCTs require intensive planning, co-ordination, expertise7 and sufficient funding. There is no national institution providing a coherent suite of support, available skills and training for clinicians wishing to conduct trials in the public sector. In many countries, clinical trials support units offer such support to clinicians and researchers in a variety of settings.8 The MRC consequently funded a feasibility study to assess the need for establishing a national South African Clinical Trials Support Unit. We report on a component of the study comprising key informant interviews of national decision-makers.



Our objective was to determine the need for additional training and support for conducting randomised clinical trials within South African institutions; identify challenges facing institutions conducting randomised clinical trials; and provide recommendations for enhancing trial conduct within South African public institutions.




The study was qualitative, comprising key informant interviews. Sampling was purposive: the research team identified senior decision-makers at stakeholder institutions; during their interviews, participants were asked to suggest additional key informants to provide contributory insights ('snow-balling'). The number of interviews was considered sufficient when participants from at least one of all the key institutions had been interviewed and no new information was forthcoming. We anticipated that a minimum of 20 interviews would be required to ensure inclusion of all universities, government organisations and non-government organisations active in the conduct, use or funding of trials.

Data collection and management

The study received ethics approval from the MRC Ethics Committee in February 2008. The principal investigator (NS) conducted the interviews between May and August 2008. Interviews were semi-structured and open-ended, made use of probes to explore key areas, and took approximately 45 minutes each. Participants were encouraged to base their answers on experience in their representative institutions and on their personal knowledge of the South African clinical trials landscape. They were audio-recorded with the interviewee's permission after signing an informed consent form. An individual unaware of the identity of the interviewee transcribed the interviews. Transcripts and identifying material were accessible only to the principal investigator.

Data analysis

The transcripts of the in-depth interviews were reviewed, and a code list of the emergent themes and sub-themes was generated for analysis. Participants were emailed the preliminary analysis for additional comment. The data are presented as frequencies, and illustrative quotations appear in italics.



Characteristics of key informants

Sixteen interviews (15 face-to-face and 1 telephonic) were conducted; 14 were conducted with a single informant, 1 was conducted with 2 informants, and 1 with 3 informants. The views of 19 senior individuals from 14 institutions were thus elicited. At 7 of the 8 South African universities with clinical faculties of health sciences, interviews were conducted at the level of a dean or equivalent. The representatives from other institutions were all at senior management levels. Two key informants who agreed to be interviewed subsequently were not able to participate in the interview. The Medicines Control Council (MCC) was invited to participate, but no representative was forthcoming.

Institutional trial experience

Of the 10 institutions which reported conducting clinical trials, 9 believed that they were experienced in trial conduct, with 2 reporting that the level of trial experience was variable across the institution. At the time of the interview, 3 institutions were conducting over 50 trials, 3 were conducting between 10 and 50 trials, and 4 were conducting less than 10 trials. Respondents whose institutions did not conduct trials reported using trial results to determine policies or were involved in administering trial research training, or funding.

Trial funding

All informants reported funding sources as broad-based, including a combination of institutional support and grants, international agencies, the MRC and pharmaceutical industry. Strong themes that emerged were: (i) lack of funding for investigator-driven (also known as self-initiated) trials; and (ii) difficulty in obtaining funding for specific research questions important to the public sector.

Strengths of and challenges to trial conduct

Participants identified the high disease prevalence and relatively strong infrastructure in the South African health system as being fundamental to potentially successful trial conduct in the country. Participants identified the lengthy MCC processes and complex logistics of RCTs, the high cost of trials and the potential burden of trial conduct on clinical care as major obstacles.

Training and support needs

Table I outlines the identified training and support needs, with research methods training a strong theme. Responses to needs for support structures were more diverse, with assistance with statistical analysis, data management and science writing emerging across the responses.

Participants were specifically asked if they believed that South Africa had sufficient and suitably experienced researchers able to conduct monitoring and quality control of RCTs. Of the 13 who expressed an opinion, 3 believed this to be true, 5 felt there were some suitably skilled researchers but insufficient for the need, and 5 stated that this was a major skills shortcoming.

Need for a clinical trials support unit

In 15 of 16 interviews, participants reported a need for an initiative at national level to encourage and improve the conduct of clinical trials. One interviewee stated that it should be focused on previously disadvantaged institutions; one stated that it should start small, and another stated that it should focus only on quality control and monitoring. Several participants expressed surprise that such an entity did not exist already.

Participants were asked to consider how such an entity would best operate and where its funding should come from. Four potential operational structures (Table II) and 4 key deliverables were identified: (i) provision of quality control, monitoring, and oversight of trials; (ii) training pertaining specifically to trials to avoid duplication with current university-based training programmes; (iii) mentoring support for the entire trial process from grant procurement to final report writing; and (iv) potential to play an advocacy role to streamline regulatory processes.

In 15 of the 16 interviews, representatives suggested that funding be derived from a national government source; specific recommendations included the MRC (5), DoH (2), Department of Education (1), Department of Science and Technology (1), national or provincial government with no departments specified (3), a combination of national departmental funding (3), and the pharmaceutical industry in partnership with the DoH (1). One respondent recommended that trial sponsors including international agencies should contribute funds for establishing such an entity.

Many participants appeared dispirited by the apparent lack of long-term sustainability of such an initiative, some citing previous failed attempts at national collaborative projects, and cautioning about the potential for competition rather than collaboration between institutions.



South African clinical trial research community senior stakeholders and decision-makers voiced the significant need for a national clinical trials initiative to support and enhance the conduct of public sector RCTs. Participants agreed that, ideally, the national government should provide specific funding to establish this initiative, and that its long-term sustainability should be carefully considered.

Kahn and Gastrow estimate that industry turnover in clinical trials run by the South African pharmaceutical industry is worth around R14.1 billion annually.9 Despite this, they argue that South Africa neither has the requisite human resources to be competitive internationally, nor does it invest enough. They report that increased trial activity could be attracted to South Africa by our well-established credentials in medical research, the high burden of disease and relatively drug-naïve populations. These factors were recognised by our participants as key reasons for conducting trials locally. Recommendations for a more proactive regulatory environment were echoed in our respondents' complaints regarding MCC waiting times and requests for assistance in negotiating the regulatory processes.

Most participants described the difficult local conditions facing trial investigators. Conducting trials within communities where literacy and health care knowledge of participants and providers is poor, has led to speculation that trials of high methodological quality are not possible within these settings.10-12 A comparison of African and North American HIV/AIDS RCTs found that the reported methodological quality of African trials was better than that of North American trials, independent of the country of residence of the principal investigator,13 which is encouraging for the future of local trial conduct.

Qualitative research was appropriate as it helps the development of concepts, giving emphasis to the meanings, experiences and views of all participants.14 Individual interviews allowed detailed exploration of the issues15 and provided rich, comprehensive data16 which might not have been achieved with a survey-driven approach. Consistency was ensured by each interview being conducted by NS. Interviews took place over a 4-month period to maximise availability of the appropriate institutional key informants. All but one of the pre-specified institutions were represented in the final sample.


The study is subject to limitations inherent in qualitative research.15 It was funded and conducted by the MRC, which has an interest in the location of such a unit, which might have influenced the investigators' interpretation of results. Participants working in government departments may be far removed from the 'coalface', and their comments may not accurately reflect needs. However, given the consistency of comments from all participants, this does not seem to have been a major limiting factor. Perhaps most limiting were the voices missing from this study, notably a stakeholder from the MCC. Feasibility issues prohibited a larger study, but opinions of more clinical investigators and of trial participants, civil society groups and community representatives would have enriched the study. Such studies could be key outputs of a future clinical trials unit.

We thank the interview participants for their time, Leon van Wyk for his excellent transcriptions, and Joy Oliver and Elizabeth Pienaar of the South African Cochrane Centre for their administrative support, assistance and enthusiasm. Nandi Siegfried acknowledges the support of the University of Cape Town's Research Office and the Department of Public Health for funding her attendance at a writers' workshop.

Conflict of interest. All authors are employees of the MRC, which funded this research.



1. Anonymous. Fifty years of randomised controlled trials. BMJ 1998; 317: 7167.         [ Links ]

2. Gevers W. Clinical research in South Africa: a core asset under pressure. Lancet 2009; 374(9692): 760-762.         [ Links ]

3. Pang T, Lansang MA, Haines A. Brain drain and health professionals. BMJ 2002; 324(7336): 499-500.         [ Links ]

4. Benatar S, Vaughan C. Global and local forces shaping the research agenda and the governance of research ethics. South African Journal of Science 2008; 104: 439-442.         [ Links ]

5. MBewu A. How academic medicine should be positioned within medicine in Africa? A global perspective. In: Golden Jubilee of the Colleges of Medicine of South Africa. Cape Town: Colleges of Medicine of South Africa, 2005.         [ Links ]

6. Department of Health. 2008/9 Annual Report. (accessed 14 October 2009).         [ Links ]

7. Schulz KF. Unbiased research and the human spirit: the challenges of randomized controlled trials. CMAJ 1995; 153(6): 783-786.         [ Links ]

8. Gluud C, Sorensen TI. New developments in the conduct and management of multi-center trials: an international review of clinical trial units. Fundam Clin Pharmacol 1995; 9(3): 284-289.         [ Links ]

9. Kahn M, Gastrow M. Pharmacologically active: clinical trials and the pharmaceutical industry. S Afr Med J 2008; 98(2): 114-116.         [ Links ]

10. Edejer TT. North-South research partnerships: the ethics of carrying out research in developing countries. BMJ 1999; 319: 438-441.         [ Links ]

11. Enabling research in developing countries (editorial). Lancet 2000; 356: 1043.         [ Links ]

12. Langer A, Diaz-Olavarrieta C, Berdichevsky K, Villar J. Why is research from developing countries underrepresented in international health literature, and what can be done about it? Bull World Health Org 2004; 82(10): 802-803.         [ Links ]

13. Siegfried N, Clarke M, Volmink J, Van der Merwe L. African HIV/AIDS trials are more likely to report adequate allocation concealment and random generation than North American trials. PloS ONE 2008; 3(10): e3491.         [ Links ]

14. Pope C, Mays N. Qualitative research: Reaching the parts other methods cannot reach: an introduction to qualitative methods in health and health services research. BMJ 1995; 311: 42-45.         [ Links ]

15. Britten N. Qualitative research: Qualitative interviews in medical research. BMJ 1995; 311: 251-253.         [ Links ]         [ Links ]



Accepted 22 January 2010.



Corresponding author: N Siegfried (

^rND^sGevers^nW^rND^sPang^nT^rND^sLansang^nMA^rND^sHaines^nA^rND^sBenatar^nS^rND^sVaughan^nC^rND^sMBewu^nA^rND^sSchulz^nKF^rND^sGluud^nC^rND^sSorensen^nTI^rND^sKahn^nM^rND^sGastrow^nM^rND^sEdejer^nTT^rND^sLanger^nA^rND^sDiaz-Olavarrieta^nC^rND^sBerdichevsky^nK^rND^sVillar^nJ^rND^sSiegfried^nN^rND^sClarke^nM^rND^sVolmink^nJ^rND^sVan der Merwe^nL^rND^sPope^nC^rND^sMays^nN^rND^sBritten^nN^rND^sMays^nN^rND^sPope^nC^rND^1A01^nM N^sSmittenberg^rND^1A02^nD^sLungelow^rND^1A03 A04^nH^sRode^rND^1A03 A05^nA B^svan As^rND^1A03 A04^nA J W^sMillar^rND^1A01^nM N^sSmittenberg^rND^1A02^nD^sLungelow^rND^1A03 A04^nH^sRode^rND^1A03 A05^nA B^svan As^rND^1A03 A04^nA J W^sMillar^rND^1A01^nM N^sSmittenberg^rND^1A02^nD^sLungelow^rND^1A03 A04^nH^sRode^rND^1A03 A05^nA B^svan As^rND^1A03 A04^nA J W^sMillar



Can fireworks-related injuries to children during festivities be prevented?



M N SmittenbergI; D LungelowII; H RodeIII; A B van AsV; A J W MillarIV

IFinal-year medical student. Erasmus University, Rotterdam, The Netherlands
IIBA. University of North Carolina, United States of America
IIIMB ChB, MMed (Surg), FRCS (Ed), FCS (SA). Red Cross War Memorial Children's Hospital, and University of Cape Town
IVMB ChB, FRCS, FRACS, DCH (RCP & S Eng), FCS. Red Cross War Memorial Children's Hospital, and University of Cape Town
VMB ChB, MMed, MBA, FCS (SA), PhD. Red Cross War Memorial Children's Hospital, and Childsafe South Africa




OBJECTIVE: To determine the epidemiological features and outcome of fireworks-related injuries among children 0 - 13 years old.
DESIGN: A retrospective study from the trauma registry of a children's hospital from 2001 - 2009.
RESULTS: Fifty-five children were treated for injuries from fireworks. The mean age was 8.8 years, 78% were boys, and the largest age group was 5 - 9 years old. Firecrackers accounted for 95% of the injuries; the most commonly injured body sites were hands (44%), eyes (42%) and face (31%); 47% of the patients had more than one injury. The most common injury type was burns (67%); 25 children were admitted, mostly to the burns and ophthalmology units. The mean length of hospital stay was 3.5 days. Surgical intervention was required in 38% of the patients. Most of the fireworks accidents occurred in or around the patients' homes. There were more fireworks-related injuries around Guy Fawkes Day (85%) than New Year's Eve (9%).
CONCLUSION: Consumer fireworks cause serious but preventable injuries to children, either as users or bystanders. Children and their families should be encouraged to enjoy pyrotechnical displays conducted by professionals at designated areas. All fireworks for individual private use should either be supervised by an adult or banned. Current legislation should be more strictly enforced, especially the sale to under-age children.



In South Africa, fireworks are widely used on New Year's Eve, the 5th of November (Guy Fawkes Day) and for Diwali celebrations (Festival of Lights) during October/November. Fireworks injury studies have found a high incidence among children. In Australia,1 88% of emergency department attendees with fireworks injuries were under 18 years of age. In the USA, 85 800 children (average age 10.8 years) were treated over a 14-year period.2 In Israel,3 during the Jewish Purim Festival and Arab and Druze Id el-Adha holidays, 53% of those injured by fireworks were children under the age of 16 years. In New Zealand, 68% of the injured were children under the age of 15 years.4

The consequences of fireworks injuries, especially among children, are not generally known.5,6 Levitz5 described 10 patients in South Africa under the age of 15 with ocular injuries due to fireworks in Johannesburg over the 1996/1997 New Year period.

Despite clear directives from local law enforcement agencies and the identification of specific areas for pyrotechnic displays, legal and illegal discharging of fireworks in public continues to occur. Our objective was to describe the epidemiology, injury pattern and outcome of fireworks-related injuries in children by using a local representative sample referred to a paediatric trauma centre, and to review fireworks-related legislation.


Material and methods

Children who sustained fireworks-related injuries were identified from the registered trauma admissions to the Red Cross War Memorial Children's Hospital, Cape Town. Fireworks/firecrackers injuries were selected for the New Year period (3 December - 14 January) and the Guy Fawkes and Diwali period (8 October - 19 November). Diwali usually falls in October/November and includes Guy Fawkes Day. A retrospective analysis for the years 2001 - 2009 was conducted, which included age, gender, nature of accident and referral. Injury types were categorised as burns, lacerations/abrasions, fractures and dislocations, without differentiating between primary and secondary injury. The parts of the body that were injured were grouped in categories: eye, face/head, hand and others. Patients were asked for details of the type of fireworks, which was recorded as firecracker or undefined fireworks, and how it was obtained. Identification of the specific type of fireworks implicated could not accurately be identified. The location of each accident was described either as own home or public place. SPSS for Windows was used to analyse data.



Injury demographics

Fifty-five children (<13 years old) with fireworks-related injuries were referred to the Red Cross War Memorial Children's Hospital trauma unit. During this period, 83 626 trauma patients were seen; fireworks cases represented 0.066%. There were 43 boys and 12 girls. Six were <4 years, 25 were 5 - 9, and 24 were >10 years old. Their mean age was 8.8 years (SD 2.7 years; range 3 - 12 years). Acquisition of the fireworks varied: across-the-counter purchases in local shops, stockpiles at home, from friends, or (allegedly) picked up in the street. Most children were unsupervised at the time of ignition. The mechanisms of fireworks injuries included the more expected, namely fireworks exploding in the hand (22) or thrown at and exploding near the injured child (32), as well as the more unusual (Table I). Most of the children were uninvolved but curious bystanders. Firecrackers were identified as the causative device in 52 cases, fireworks in 2, and 1 child was forced to swallow gunpowder. No homemade fireworks were identified. Injuries were sustained either at home or unintentionally at public celebrations, e.g. New Year's Eve at Clifton beach, but none at an official pyrotechnics display. Forty of the injuries occurred at home, 4 at public places, and 11 at unknown sites.

Most injuries occurred around Guy Fawkes Day (47), 3 during the period of Diwali celebrations, and 5 at New Year's Eve. Table I depicts typical scenarios preceding the injury.

Anatomical considerations

One anatomical site only was involved in 53% of the injuries. The most frequent types of injury were burns, lacerations and abrasions (Table II). The most common injuries were to the hands, with partial thickness burns; lacerations; dislocations; amputation of fingers; and fractures. Three children sustained body surface burns of 2 - 28% needing extensive skin grafting. The eyes were the next most common site, with corneal abrasions/lacerations, hyphaema, foreign bodies and one eye requiring enucleation for globe disruption. Eye injuries are potentially very serious, and included sphincter tears, traumatic uveitis, choroidal rupture adjacent to the fovea, and 3 pre-retinal haemorrhages in the temporal macula. Faces sustained flash burns and lacerations, with burns and lacerations to other body parts. One child hid a burning firework in his pants, resulting in a pelvic fracture, genital ablation, necrotic anterior abdominal wall, perforated bowel and degloving of the femoral triangle. A small cracker was inserted into the auditory canal of one child and lit, resulting in perforation and deafness. Influenced by her friends, one 3-yearold ingested gunpowder. Firecracker remnants were found in 2 hands and 2 eyes and were surgically removed. Gunpowder staining in 7 wounds was dissolved and removed following the topical application of liquid paraffin.



The mean age of the 25 children admitted was 9.4 years (SD 2.9 years, range 3 - 12 years). Reasons for admission were injuries to the hands, face and eye. Placement included admission to the trauma ward, burns unit and ophthalmology ward, and one child was admitted to the ICU. Surgical intervention for the injuries was required in 21 cases. Information about duration of hospital stay was available in 80% of cases. The mean length of hospitalisation was 3.5±3.2 days. None of the patients seen died from their injuries.



Despite the annual outcry over the indiscriminate use of fireworks during festive periods and the concomitant cruelty to animals, very little is known or has been published about injuries sustained by humans in South Africa.5,6 Children, because of their inquisitive nature and the excitement surrounding fireworks, are particularly vulnerable. They also cannot readily appreciate the dangers involved or act decisively in an emergency, hence their susceptibility to firework injuries. The fireworks responsible were freely available and often sold to under-age children and then indiscriminately used. Injuries were more frequent and more severe among children who were active participants rather than bystanders. Most self-inflicted injuries occurred among boys, whereas girls suffered injuries mostly as bystanders.

In this study, the accidents occurred mainly in or around the home and not at public places, which emphasises the ignorance or lack of consideration about the dangers, or the need for supervision. On festive days, fireworks were most often not used in allocated discharging sites. Although we know that firecrackers (sparklers, novelty designs, penny bangs, rockets, etc.) are dangerous, no information or instructions are available to the public regarding the explosive capacity, time delay from ignition to explosion, the wick, and effects of its combustion. Misuse of the devices can also lead to unintentional collateral damage to property, which was not reported in this survey.

The main 'culprits' are firecrackers and sparklers, which contain up to 130 g of powder charge and may generate temperatures exceeding 500ºC.7 The danger of firecrackers is the power and unpredictable timing of the gunpowder explosion; they should be classified as explosives.7 Firecracker/ rocket injuries primarily involve the hands and eyes, while sparklers cause burns either directly or through setting clothing alight.3,7-9 Malfunctioning devices account for only a small percentage of injuries.9

Although only 45 patients were admitted to the trauma centre, the consequences for some children were significant. Fortunately, most injuries were not severe and, after treatment, children could be released from the emergency department. However, many of those admitted required surgery. The most commonly injured body parts were hands (non-dominant), eyes and face, as mentioned in other reports.5,7-11 As may be expected from the combustion and heat generated, burns accounted for most of the injuries, leading to loss of fingers, visual impairment, deafness, lacerations, abrasions, contusions and permanent facial and body scars.6,8,9 Nearly half (47%) of the patients had more than one injury. Burn wounds were all of a contact type and of partial depth; all but 3 healed with topical therapy.

There was a preponderance of coloured and white children with fireworks-related injuries; only 4 of the injured children were black, which could be a reflection of different cultural customs.

Studies emphasise the high incidence of fireworks-related injuries during holiday and religious periods, and are a global concern. In the USA, 95% of such injuries occur around Independence Day; in Greece, there is a peak during the Greek Orthodox Easter; in Iran during their end-of-year celebrations; and in India during the Diwali Festival.8,9,12 No injuries occurred during the Halloween period.

Fireworks in South Africa are regulated by the Explosives Act (Act 26 of 1956) which states that the police grant licences to sell fireworks, and a retailer may not sell fireworks to children under the age of 16 years.13 Sale of fireworks by informal traders/street hawkers is illegal. Fireworks may only be sold in sealed packages, as received from the supplier. Importers of fireworks require a permit to do so.13 A child under the age of 16 years may not handle or use fireworks except when supervised by an adult.

Regionally, specific municipal by-laws regulate the use of fireworks. In Cape Town, discharging fireworks in public is illegal, without obtaining prior permission from the Chief Inspector of Explosives and the City of Cape Town.13 The fine for committing this offence is R600 or 12 months in prison. However, the City permits the discharging of fireworks in public on Guy Fawkes Day and New Year's Eve, but only in specially allocated sites (by-law 58:12). In 2009, the City of Cape Town identified 19 public areas that could be used for Diwali and for Guy Fawkes.13 Some municipalities do not require their residents to apply for permission for using fireworks for religious and cultural celebrations, and allow them to be used at home.

Despite legislation, legal practices are often evaded. According to the amendment inserted in Chapter 11, section 58, by-laws were violated in many cases in our study,13 including: 58:2 - restriction of sale of fireworks by licensed shops, 58:10 - restriction of firework sales exclusively to permit holders, 58:4 - the requirement for an operational permit, and by-law 58:12 - designation of approved firework discharging sites. Furthermore, illegal use of fireworks can be reported to the metropolitan police and the culprits be fined if they cause a disturbance, nuisance or damages or danger to anyone.

To prevent fireworks-related injuries among children, a comprehensive approach is necessary.2,4,7-10 Recommendations include parental education and supervision at home during festive seasons, and only attending formally organised fireworks displays.8,13 Legislation has a significant effect on injury rates. In the USA, the 'injury rate was seven times greater in states with fewer restrictions'.8 In Hungary, incidences were reduced following a 'legislative ban on private fireworks displays'.9

Many studies conclusively recommend the introduction of stricter regulations and banning the indiscriminate sale of fireworks. There should also be greater control of retailers, the public be encouraged to use pre-approved firework sites, and school education initiatives be supported.15 Proposed amendments to by-laws include that fireworks can only be used by the permit holder, for a certain time duration, and in the 'premises for which it was issued'. As children cannot obtain permits, they would be prevented from handling fireworks. However, these terms would not apply to Guy Fawkes and New Year's Day, for which occasions designated areas are determined for public use and are 'subject to conditions as may be determined by the controlling authority'.13 As the greatest fireworks use and highest record of child injury is on festive days, by-laws should also apply to these days.

Limitations of our study were the retrospective nature and that it might not be representative of all children injured during the study period. A review of all emergeny room (ER) cases of all hospitals in Cape Town would be required before drawing conclusions regarding the prevalence, especially around Guy Fawkes Day. The mid-year population estimate for the Western Cape in 2009 was 5 356 900 people. Since 2005, no fatality has been reported from the Department of Forensic Pathology at the University of Cape Town.

In conclusion, fireworks are associated with serious but preventable injuries among the paediatric age group. Parents should take their children to safer public fireworks displays rather than allow consumer fireworks to be used by or near their children.16 More enforcement of regulations, education and parental supervision are needed to reduce fireworks-related injuries, and mail-order pyrotechnics should be banned.



1. Abdulwadud O, Ozanne-Smith J. Injuries associated with fireworks in Victoria: an epidemiological overview. Inj Prev 1998; 4(4): 272-275.         [ Links ]

2. Witsaman RJ, Comstock D, Smith GA. Pediatric fireworks-related injuries in the United States: 1990-2003. Pediatrics 2006; 118: 296-303.         [ Links ]

3. Zohar Z, Waksman I, Stolero J, Volpin G, Sacagiu E, Eytan A. Injury from fireworks and firecrackers during holidays. IMA 2004; 143: 698-701.         [ Links ]

4. Clarke J, Langley JD. Firework related injury in New Zealand. NZ Med J 1994; 107: 423-425.         [ Links ]

5. Levitz LM. Children and fireworks - a dangerous combination. S Afr Med J 1997; 87(11): 1554-1556.         [ Links ]

6. Wilson V. Firework injuries. Trauma and Emergency Medicine 1999; 16(2): 19-26.         [ Links ]

7. Berger LR, Kalishman S, Rivara FP. Injuries from fireworks. Pediatrics 1985: 75(5): 877-882.         [ Links ]

8. Smith GA, Knapp JF, Barnett TM, Shields BJ. The rockets' red glare, the bombs bursting in air: Fireworks-related injuries to children. Pediatrics 1996; 98(1): 1-9.         [ Links ]

9. Kuhn FC, Morris RC, Witherspoon CD, et al. Serious fireworks-related eye injuries. Ophthalmic Epidemiol 2000; 7(2): 139-148.         [ Links ]

10. Committee on Injury and Poison Prevention. Fireworks-related injuries to children. Pediatrics 2001; 108: 190-191.         [ Links ]

11. Arya S, Malhotra S, Dhir S, Sood S. Ocular fireworks injuries. Clinical features and visual outcome. Ind J Oph 2001; 49(3): 189-190.         [ Links ]

12. Vassilia K, Eleni P, Dimitrios T. Firework-related childhood injuries in Greece: a national problem. Burns 2004; 30: 151-153.         [ Links ]

13. Provinsie Wes-Kaap. Provinsiale Koerant 6347. Chapter 11 - Fireworks. 3 Maart 2006.         [ Links ]

14. City appeals to residents to ensure an accident-free Guy Fawkes. Cape Gateway 29 October 2009. (accessed 1 November 2009).         [ Links ]

15. Sheller JP, Muchardt O, Jonsson B, Mikkelsen MB. Burn injuries caused by fireworks: effect of prophylaxis. Burns 1995; 21(1): 50-53.         [ Links ]

16. Childsafe. Play it safe with fireworks. (accessed 1 November 2009).         [ Links ]



Accepted 4 February 2010.



Corresponding author: H Rode (

^rND^sAbdulwadud^nO^rND^sOzanne-Smith^nJ^rND^sWitsaman^nRJ^rND^sComstock^nD^rND^sSmith^nGA^rND^sZohar^nZ^rND^sWaksman^nI^rND^sStolero^nJ^rND^sVolpin^nG^rND^sSacagiu^nE^rND^sEytan^nA^rND^sClarke^nJ^rND^sLangley^nJD^rND^sLevitz^nLM^rND^sWilson^nV^rND^sBerger^nLR^rND^sKalishman^nS^rND^sRivara^nFP^rND^sSmith^nGA^rND^sKnapp^nJF^rND^sBarnett^nTM^rND^sShields^nBJ^rND^sKuhn^nFC^rND^sMorris^nRC^rND^sWitherspoon^nCD^rND^sArya^nS^rND^sMalhotra^nS^rND^sDhir^nS^rND^sSood^nS^rND^sVassilia^nK^rND^sEleni^nP^rND^sDimitrios^nT^rND^sSheller^nJP^rND^sMuchardt^nO^rND^sJonsson^nB^rND^sMikkelsen^nMB^rND^1A01^nIan Paul^sOlwoch^rND^1A01^nIan Paul^sOlwoch^rND^1A01^nIan Paul^sOlwoch



The microbiology of acute complicated bacterial sinusitis at the University of the Witwatersrand



Ian Paul Olwoch

MB ChB, MMed (Anaesth), MMed (Otorhinolaryngol). Department of Otorhinolaryngology, Head and Neck Surgery, University of the Witwatersrand, Johannesburg




DESIGN: A retrospective chart review study at two referral hospitals identified 226 consecutive surgical patients with acute complicated sinusitis.
SUBJECTS: One hundred and fifty-nine male and 67 female patients, with a mean age of 16.5 (standard deviation 0.7) years, underwent external fronto-ethmoidectomy with maxillary sinus washout and 13 had a concurrent craniotomy.
RESULTS: A total of 233 micro-organisms were isolated from 163 patients (72.1%), and 63 (27.9%) were culture negative. Positive isolates included Streptococcus milleri (18.5%), Staphylococcus aureus (12.4%),
β-haemolytic streptococci (10.8%), coagulase-negative staphylococci (8.6%), Haemophilus influenzae (8.6%) and the anaerobes, Peptostreptococcus (6.4%) and Prevotella (4.7%) species. The prevalences of S. pneumoniae (2.6%), methicillin-resistant S. aureus (MRSA) (1.3%) and Moraxella catarrhalis (0.4%) were low. Polymicrobial disease was present in 56 patients (34.4%). There was a significant difference between the two hospitals in the prevalences of some bacteria (p<0.05).
Antibiotic resistance was highest towards the penicillins (64.3%) and cephalosporins (12.5%). Effective empiric treatment was achieved with metronidazole and a choice of amoxicillin-clavulanate or ampicillin plus cloxacillin or penicillin plus chloramphenicol.
CONCLUSION: The polymicrobial nature and severity of complicated sinusitis warrants a de-escalation approach to antimicrobial therapy. The combination of
β-lactamase- resistant penicillins and metronidazole is a reasonable choice for initial empiric antibacterial therapy.
Selection of drugs for empirical antibiotic therapy in patients with acute complicated sinusitis should be supported by knowledge of the local prevalence and antimicrobial susceptibilities of bacteria isolated from patients.



The microbiology of bacterial sinusitis has been studied extensively. The most common pathogens cited are aerobes and facultative anaerobes that include Streptococcus pneumoniae, Haemophilus influenzae, Staphylococcus aureus, Moraxella catarrhalis, Klebsiella pneumoniae, Pseudomonas aeruginosa and S. milleri.1- 9 Anaerobic organisms, isolated from as many as one-third of patients, include Propionibacterium acnes and Peptostreptococcus, Prevotella, Enterobacter, Bacteroides and Fusobacterium species.9,10 Despite the abundance of published studies there are very few that report on the microbiology of acute complicated sinusitis.3-8,11,12

Complicated sinusitis is associated with debilitating and potentially life-threatening orbital, intracranial, bone and soft-tissue complications; its treatment is both a medical and a surgical emergency.3,5,6,11,13,14 At the University of the Witwatersrand, patients with complicated sinusitis are referred to the Department of Otorhinolaryngology where they usually present with any combination of sinus involvement (maxillary, ethmoid, frontal and sphenoid) and one or more associated complications.3,5,6,11,13,14 A routine clinical assessment and computed tomography scan diagnosis is made before any surgery. Prompt medical treatment, with or without surgery, is then commenced. This entails initial empirical use of antibiotics before laboratory identification of the offending pathogens. As in other centres, observed differences in the local prevalence of bacterial pathogens and their antibiotic susceptibilities will influence the choice of drugs used at the start of treatment.1,3,4,6,7

The aims of this study were to determine the bacteriology and antibiotic susceptibility in isolates from patients with complicated sinusitis at the University of the Witwatersrand and to discuss an empirical basis for the initial choice of antibiotic drugs at this institution.


Materials and methods

A 5-year (January 2002 - December 2006) retrospective hospital chart review of patients with acute complicated sinusitis who underwent external fronto-ethmoidectomy (or ethmoidectomy) with maxillary sinus washout was performed. Institutional approval was obtained for the study (Protocol No. M070203). Patients were identified from the registers of the operating theatre, surgical wards and radiology department in each of two university hospitals situated 30 km apart: Johannesburg (JH) and Chris Hani Baragwanath (CHB) hospitals. The database of the resident National Health Laboratory Service (NHLS) was searched for corresponding records of microbiological investigations performed on sinus specimens derived at the time of surgery.

Sinus aspirate and tissue specimens were collected intra operatively from affected sinuses, and standard aseptic conditions of collection were presumed to have been applied. The specimens were then transported to the laboratory, usually within 2 hours of collection. Sinus aspirates and ground tissues were inoculated onto 5% horse blood agar, MacConkey agar, chocolate agar and thioglycolate broth for aerobic bacteria and incubated at 35ºC for 24 - 48 hours. Specimens for anaerobic culture were obtained directly from the surgical samples and indirectly as a subculture from thioglycate broth. Anaerobic specimens were inoculated onto 10% horse blood agar and amikacin blood agar and incubated at 35ºC for 24 - 48 hours. Direct Gram staining and microscopy were first performed after 24 hours of incubation and then repeated as necessary at 48, 72 and 96 hours.

The predominant pathogens isolated in the microbiology cultures were used for statistical analysis. The following variables were analysed: (i) gender; (ii) age (paediatric 0 - 15 years or adult >15 years); (iii) presence (or absence) of intracranial complications; (iv) number of different species (monomicrobial or polymicrobial) found within a single specimen; and (v) hospital from which the sample was obtained.

Only cases with positive sinus microbiological cultures were included in the evaluation. Statistical analysis was performed using Student's t-test and the chi-square (χ2) test. A p-value of <0.05 was regarded as significant.



A total of 301 surgical patients were identified; 75 were excluded from the study owing to an absence of laboratory records (42), sinus histology reports without corresponding microbiology reports (29), incomplete microbiology reports (3) and rejected specimens (1).

Microbiology culture results were obtained for 226 patients, 159 males and 67 females (mean age 16.5 (standard deviation (SD) 0.7) years). All underwent external frontoethmoidectomy with maxillary sinus washout and 13 had a concurrent craniotomy. One hundred and sixty-three specimens (72.1%) were culture positive and 63 (27.9%) culture negative. A total of 233 micro-organisms were isolated from the culture-positive specimens (Table I). Aerobes and facultative anaerobes accounted for 85.5% of the isolates, anaerobes for 13.3% and fungi for 1.2%. The most frequently isolated aerobic micro-organisms were S. milleri (18.5%), S. aureus (12.4%), β-haemolytic streptococci (10.3%), coagulase-negative staphylococci (8.6%) and H. influenzae (8.6%). S. pneumoniae (2.6%) and M. catarrhalis (0.4%) were not common pathogens. The prevalence of methicillin-resistant S. aureus (MRSA) was very low (1.3%). The anaerobes isolated were mainly Peptostreptococcus (6.4%) and Prevotella (4.7%) spp.



Comparison by gender (Table II). Males (71.8%) were affected more often than females (28.2%), with a male/female ratio of 2.7:1. Of the 31 patients with intracranial complications, 29 males and 3 females were affected (data not shown). No significant differences in the occurrence of bacterial species were observed within the gender study group.

Comparison of paediatric and adult patients (Table II). There were 70 adults (mean age 24.2 (SD 1.3) years) and 93 children (mean age 10.8 (SD 0.3) years) with culture-positive specimens. H. influenzae was isolated more frequently from the paediatric group (p<0.05) and coagulase-negative staphylococci mainly from adults (p<0.05).

Comparison by presence or absence of intracranial complications (Table II). Patients with and without intracranial complications were affected by the same species of bacteria and to a similar extent. S. milleri was the most common microorganism and was isolated from 48.4% and 21.2% of patients with and without complications, respectively.

Comparison of monomicrobial and polymicrobial specimens (data not shown). Isolates from 65.6% of the culture-positive specimens were monomicrobial, while 34.4% were polymicrobial. The mean number of species recovered per specimen was 1.4 (range 1 - 4). One specimen contained four species, namely S. milleri, S. viridans, H. influenzae and Photobacter species. A significantly higher (p<0.05) percentage of aerobic and facultative anaerobic bacteria (95.3%) were isolated from monomicrobial specimens compared with polymicrobial specimens (75.4%). Anaerobes were isolated in 4.7% and 24.6% of monomicrobial and polymicrobial specimens, respectively (p<0.05). Peptostreptococcus and Prevotella spp. were isolated exclusively from polymicrobial specimens.

Comparison of specimens from JH and CHB (Table II). A similar study population of patients was found at each of the two hospitals. Sixty-nine culture-positive patients (42.3%) aged 3 - 48 years (mean 18.4 (SD 1.2) years) were sourced from JH and 94 patients, aged 4 - 74 years (mean 15.3 (SD 1.0) years), from CHB. Eighty-seven and 146 micro-organisms were isolated from JH and CHB, respectively. S. milleri was the most common isolate at CHB (28.1%), where it occurred more frequently (p<0.05) than at JH (2.3%). Beta-haemolytic streptococci were the most common micro-organisms found at JH (20.7%) and were a more frequent (p<0.05) isolate than at CHB (4.8%). Coagulase-negative staphylococci were also a more frequent (p<0.05) pathogen at JH (14.9%) than at CHB (4.8%). There was no significant difference between the two hospitals in the reported frequencies of S. aureus and Peptostreptococcus and Prevotella species.

Antibiotic susceptibility (data not shown). Records were obtained for antibiotic susceptibility for 78 microorganisms; however, susceptibility tests were not consistently reported for all common antibiotics, and significance levels were therefore not tested. Penicillin, ampicillin and erythromycin were highly effective against S. milleri, β-haemolytic streptococci, S. pneumoniae and streptococcal species. S. aureus and coagulase-negative staphylococci were highly susceptible to cloxacillin and resistant to penicillin. Coagulase-negative staphylococci showed no resistance to either ampicillin or erythromycin. H. influenzae was susceptible to ampicillin in 7 of 9 cases. MRSA was susceptible to vancomycin.

Antibiotic resistance (Table III). One hundred and twelve instances of antibiotic resistance were recorded. The highest occurrence of resistance was to the penicillins (64.3%), cephalosporins (12.5%) and co-trimoxazole (8%). There were 4 reports of resistance of anaerobic bacteria to clindamycin, and no susceptibility tests to metronidazole were documented.




Initial empiric antibiotic therapy in patients with acute complicated sinusitis should ideally be guided by the knowledge of the most prevalent causative pathogens and their antimicrobial resistance patterns. The prevalence of pathogenic bacteria and antimicrobial resistance patterns may vary widely between different centres located within the same geographical region.1,4-6 It is therefore important for medical practitioners to be aware of the existing trends in the communities in which they practise. In medical literature the aerobic species cited as being among the most prevalent include S. aureus,3,6,7,11-13 H. influenzae,1,2,6,9 S. milleri,4,5 S. pneumoniae,2,9,12 S. viridans3,6 and M. catarrhalis.2,9 The most frequently reported anaerobes include Peptostreptococcus, Prevotella, Fusobacterium, Propionibacterium and Bacteroides.5,9-11

In this study 226 cultures were recovered, predominantly from young male patients. Of these 163 (72.1%) were culture positive and yielded a total of 233 isolates. Negative culture rates are a measure of prior antibiotic consumption, specimen handling and laboratory standards, and may vary widely from 0% to above 50%.6,12 In this study, the negative culture rate of 27.9% is comparable to that encountered in similar studies.2,4,11

Recent publications mention S. milleri as being the most common pathogen in patients with complicated sinusitis.4,5,8 Whereas this is in agreement with the overall findings of this study, it was true of CHB (28.1%) but not of JH (2.3%). The notable absence of S. milleri as seen at JH is also reported by other authors.3,6,7,11-13

Lancefield group A, C, F and G β-haemolytic streptococci (10.8%) were isolated at both hospitals. Group C, with 11 isolates, had the highest count, followed by groups A and F, then finally by group G (Table I). As a group they were the most common isolate at JH (20.7%) but not at CHB (4.8%). Few studies on complicated sinusitis make specific mention of β-haemolytic streptococci as a pathogen.1,7,15 In one recent study the authors present what they consider to be the first reported case of the isolation of group C streptococcus, in complicated sinusitis, from a child with an orbital sub-periosteal abscess.15

In this study S. aureus (12.4%) was the second most common isolate. This is in keeping with the universal acceptance of this organism as a major pathogen.3-8,11,14,15

H. influenzae (8.6%) and coagulase-negative staphylococci (8.6%) ranked joint third. H. influenzae occurred more frequently in children than in adults (Table II). Previous studies cite wide variations, of between 2.8% and 26%, in the incidence of positive isolates of H. influenzae in complicated sinusitis.1,2,4-6,12 Equivalent studies attach no importance to H. influenzae as a pathogen.3,8,11,13

Coagulase-negative staphylococci have long been regarded as non-pathogenic, but in a recent study their role as important pathogens is recognised.6 In this study coagulase-negative staphylococci occurred more frequently in adults than in children (Table II).

S. pneumoniae (2.6%) was infrequent and was not isolated from any patient with intracranial complications in this series. Reports from several smaller studies quote higher isolation rates, ranging from 7.7% to 11.9%.1,2,4,6,8,12 On the other hand, total absence of S. pneumoniae is also reported.7

Studies have shown M. catarrhalis to be a frequent isolate in cultures from patients with acute sinusitis.2,9 This study identified only one isolate of M. catarrhalis (0.4%), which supports the consensus that this organism is not a major pathogen in complicated sinusitis.1,3- 6,11-14

Anaerobic bacteria may account for as much as one-third of all isolates.5,6 In this study they accounted for 13.3% (Table I). In patients with intracranial complications a higher incidence (20.8%) was observed. The anaerobic isolates consisted of a similar variety of micro-organisms to that reported in other studies.7-9,12 The most frequently isolated anaerobes were Peptostreptococcus (6.4%) and Prevotella species (4.7%). Some authors report an absence of anaerobes that is in part attributed to lack of anaerobic culture tests.2,4,11

Fifty-six out of 163 (34.4%) positive culture specimens were polymicrobial (Table II). Similar findings are reported in other studies.3 The implications for therapy are that bacteria with different antimicrobial resistance are frequently found within the same patient. This is illustrated in Table II by the co-existence of S. milleri, which is highly susceptible to penicillin, with penicillin-resistant micro-organisms in 18 out of 43 isolates. Polymicrobial cultures, along with the high prevalence of S. aureus and coagulase-negative staphylococci, strongly suggest that acute exacerbation of chronic sinusitis is the cause of the disease in the affected patients.9

In addition to other essential treatments, patients presenting with complications must receive intravenous antimicrobials that cover the most likely pathogens. As shown in this and other studies, the causative pathogens (Table II) and, in all likelihood, resistance patterns vary from one hospital to another.1,4-6 Recommendations for initial therapy should therefore be based on an up-to-date hospital database of pathogens and resistance patterns. Two therapeutic approaches are described; a conservative 'traditional' approach and a more radical 'de-escalation' approach. Both regimens are modified on the basis of subsequent microbiology reports. The traditional approach advocates the use of a narrow-spectrum antibiotic, with broad-spectrum drugs held in reserve for patients who are severely ill, do not respond to treatment or are immunocompromised.4 The de-escalation approach promotes the use of broad-spectrum antibiotics.1,3,5,11 In one published report the authors recommended the empiric use of a combination of vancomycin, a third- or fourth-generation cephalosporin, and metronidazole or clindamycin.3

At JH and CHB a de-escalation approach is used, with drugs that are effective against both β-lactamase-positive aerobic bacteria and anaerobes.1,4,8,11 For aerobic bacteria this typically includes either amoxicillin plus clavulanate or ampicillin plus cloxacillin. Erythromycin is used in patients with penicillin sensitivity and cephalosporins are held in reserve. Anaerobic cover is provided by metronidazole. In patients with intracranial complications a combination of penicillin, chloramphenicol and metronidazole is often used. The low prevalence of MRSA (1.3%) in this study does not warrant the empiric use of vancomycin.



Selection of drugs for empiric antibiotic therapy in patients with acute complicated sinusitis should be supported by knowledge of the prevalence and antimicrobial susceptibilities of bacteria isolated from patients.

S. milleri, S. aureus, β-haemolytic streptococci, coagulasenegative staphylococci, H. influenzae and the anaerobes Peptostreptococcus and Prevotella species are common causative pathogens of acute complicated bacterial sinusitis at the two hospitals studied. H. influenzae was a more frequent pathogen in children. S. pneumoniae and M. catarrhalis did not appear to be major pathogens. A significant difference in the prevalence of major pathogenic bacteria between the two hospitals is demonstrated and highlights the need for medical practitioners to be aware of the trends in their own locality.

The high incidence of polymicrobial disease (34.4%) coupled with the severity of complicated sinusitis present an argument in favour of a de-escalation approach to antimicrobial therapy. A choice of metronidazole in combination with either amoxicillin-clavulanate or ampicillin plus cloxacillin provides sufficient antimicrobial cover against the locally prevalent micro-organisms at JH and CHB.

Similar studies should be performed on a regular basis at hospitals involved in the management of acute complicated sinusitis. Continuous monitoring of the prevalence and detection of the emergence of antimicrobial-resistant bacteria will guide hospital treatment policy.

I acknowledge Professor Pradip Modi for his advice on the study design and data analysis, Dr Razwi Ahmed for facilitating data collection, and Professor James Loock for his constructive editorial comments during the final preparation of the manuscript.



1. Snyman J, Classen AJ, Botha PL. A microbiological study of acute maxillary sinusitis in Bloemfontein. S Afr Med J 1988; 74: 444-445.         [ Links ]

2. Tellez I, Alba LMD, Reyes MG, Patton E, Hesles HdlG. Microbiology of acute sinusitis in Mexican patients. Arch Med Res 2006; 37: 395-398.         [ Links ]

3. Glickstein JS, Chandra RK, Thompson JW. Intracranial complications of pediatric sinusitis. Otolaryngol Head Neck Surg 2006; 134 (5): 733-736.         [ Links ]

4. Mortimore S, Wormald PJ, Oliver S. Antibiotic choice in acute and complicated sinusitis. J Laryngol Otol 1998; 112: 264-268.         [ Links ]

5. Oxford LE, McClay J. Complications of acute sinusitis in children. Otolaryngol Head Neck Surg 2005; 133: 32-37.         [ Links ]

6. Tshifularo M, Monama GM. Complications of inflammatory sinusitis in children: institutional review. South African Family Practice 2006; 48: 16-19.         [ Links ]

7. Brook I. Microbiology and antimicrobial treatment of orbital and intracranial complications of sinusitis in children and their management. Int J Pediatr Otorhinolaryngol 2009; 73: 1183-1186.         [ Links ]

8. Kombogiorgas D, Seth R, Athwal R, Modha J, Singh J. Suppurative intracranial complications of sinusitis in adolescence. Single institute experience and review of literature. Br J Neurosurg 2007; 21: 603-609.         [ Links ]

9. Brook I, Foote PA, Frazier EH. Microbiology of acute exacerbation of chronic sinusitis. Ann Otol Rhinol Laryngol 2005; 114: 573-576.         [ Links ]

10. Brook I. Bacteriology of acute and chronic ethmoid sinusitis. J Clin Microbiol 2005; 43(7): 3479-3480.         [ Links ]

11. Ali A, Kurien M, Mathews SS, Mathew J. Complications of infective rhinosinusitis: experience from a developing country. Singapore Med J 2005; 46: 540-544.         [ Links ]

12. Brook I. Microbiology of intracranial abcesses and their associated sinusitis. Arch Otolaryngol Head Neck Surg 2005; 131: 1017-1019.         [ Links ]

13. Rosenfeld EA, Rowley AH. Infectious intracranial complications of sinusitis, other than meningitis, in children: 12 year review. Clin Infect Dis 1994; 18: 750-754.         [ Links ]

14. Mortimore S, Wormald PJ. The Groote Schuur hospital classification of the orbital complications of sinusitis. J Laryngol Otol 1997; 111: 719-723.         [ Links ]

15. Chin TO, Syrimi M, Ne SC, Twomey J. Group C Streptococcal subperiosteal abscess of the orbit. Orbit 2009; 28: 160-161.         [ Links ]



Accepted 3 March 2010.



Corresponding author: I P Olwoch (

^rND^sSnyman^nJ^rND^sClassen^nAJ^rND^sBotha^nPL^rND^sTellez^nI^rND^sAlba^nLMD^rND^sReyes^nMG^rND^sPatton^nE^rND^sHesles^nHdlG^rND^sGlickstein^nJS^rND^sChandra^nRK^rND^sThompson^nJW^rND^sMortimore^nS^rND^sWormald^nPJ^rND^sOliver^nS^rND^sOxford^nLE^rND^sMcClay^nJ^rND^sTshifularo^nM^rND^sMonama^nGM^rND^sBrook^nI^rND^sKombogiorgas^nD^rND^sSeth^nR^rND^sAthwal^nR^rND^sModha^nJ^rND^sSingh^nJ^rND^sBrook^nI^rND^sFoote^nPA^rND^sFrazier^nEH^rND^sBrook^nI^rND^sAli^nA^rND^sKurien^nM^rND^sMathews^nSS^rND^sMathew^nJ^rND^sBrook^nI^rND^sRosenfeld^nEA^rND^sRowley^nAH^rND^sMortimore^nS^rND^sWormald^nPJ^rND^sChin^nTO^rND^sSyrimi^nM^rND^sNe^nSC^rND^sTwomey^nJ^rND^nA^sKoopowitz^rND^1A01 A02^nP^sSmith^rND^1A01 A02^nN^svan Rensburg^rND^1A01^nA^sRudman

Creative Commons License All the contents of this journal, except where otherwise noted, is licensed under a Creative Commons Attribution License