RESEARCH

"Other patients are really in need of medical attention"– the quality of health services for rape survivors in South Africa

« D'autres patientes ont réellement besoin d'une attention médicale » - qualité des services de santé pour les victimes de viols en Afrique du Sud

«Hay otros pacientes que realmente necesitan atención médica» - Calidad de los servicios de salud para supervivientes de violaciones en Sudáfrica

Nicola J. Christofides^I,II,1; Rachel K. Jewkes^I; Naomi Webster^I; Loveday Penn-Kekana^III; Naeema Abrahams^I; Lorna J. Martin^IV

^IMedical Research Council, Gender and Health Group, Private Bag X385, Pretoria 0001, South Africa
^IIWomen's Health Project, School of Public Health, University of Witwatersrand, Witwatersrand, South Africa
^IIICentre for Health Policy, School of Public Health, University of Witwatersrand, Witwatersrand, South Africa
^IVDivision of Forensic Medicine and Toxicology, University of Cape Town, Cape Town, South Africa

ABSTRACT

OBJECTIVE: To investigate in the South African public health sector where the best services for rape survivors were provided, who provided them, what the providers' attitudes were towards women who had been raped and whether there were problems in delivering care for rape survivors.
METHODS: A cross-sectional study of facilities was carried out. Two district hospitals, a regional hospital and a tertiary hospital (where available) were randomly sampled in each of the nine provinces in South Africa. At each hospital, senior staff identified two doctors and two nurses who regularly provided care for women who had been raped. These doctors and nurses were interviewed using a questionnaire with both open-ended and closed questions. We interviewed 124 providers in 31 hospitals. A checklist that indicated what facilities were available for rape survivors was also completed for each hospital.
FINDINGS: A total of 32.6% of health workers in hospitals did not consider rape to be a serious medical condition. The mean number of rape survivors seen in the previous six months at each hospital was 27.9 (range = 9.3–46.5). A total of 30.3% of providers had received training in caring for rape survivors. More than three-quarters of regional hospitals (76.9%) had a private exam room designated for use in caring for rape survivors. Multiple regression analysis of practitioner factors associated with better quality of clinical care found these to be a practitioner being older than 40 years (parameter estimate = 2.4; 95% confidence interval (CI) = 0.7–5), having cared for a higher number of rape survivors before (parameter estimate = 0.02; 95% CI = 0.001–0.03), working in a facility that had a clinical management protocol for caring for rape survivors (parameter estimate = 2; 95% CI = 0.12–3.94), having worked for less time in the facility (parameter estimate = -0.2; 95% CI = -0.3 to -0.04) and perceiving rape to be a serious medical problem (parameter estimate = 2.8; 95% CI = 1.9–3.8).
CONCLUSION: There are many weaknesses in services for rape survivors in South Africa. Our findings suggest that care can be improved by disseminating clinical management guidelines and ensuring that care is provided by motivated providers who are designated to care for survivors.

Keywords: Rape/rehabilitation; Health services; Quality of health care; Attitude of health personnel; Forensic medicine; Health care surveys; Cross-sectional studies; South Africa (source: MeSH, NLM).

RÉSUMÉ

OBJECTIF: Notre objectif était de déterminer, dans le secteur de la santé publique en Afrique du Sud, où sont dispensés les meilleurs services pour les victimes de viols, qui dispense ces services, quelle a été l'attitude des soignants à l'égard des femmes violées et si la prestation de soins aux victimes de viols pose des problèmes.
MÉTHODES: Une étude transversale des services a été effectuée. Deux hôpitaux de district, un hôpital régional et un hôpital tertiaire (le cas échéant), ont été choisi au hasard dans chacune des neuf provinces de l'Afrique du Sud. Dans chaque hôpital, la direction a désigné deux médecins et deux infirmières régulièrement appelés à soigner des femmes victimes de viols. Un questionnaire comprenant des questions ouvertes et des questions fermées a été utilisé pour interroger ces médecins et ces infirmières. Nous avons interrogé 124 dispensateurs de soins dans 31 hôpitaux. Une liste de contrôle indiquant les services disponibles pour les victimes de viols a également été remplie pour chaque hôpital.
RÉSULTATS: Pour 32,6% des agents de santé des hôpitaux, le viol était sans gravité au plan médical. Le nombre moyen de victimes de viols examinées au cours des six mois écoulés dans chaque hôpital était de 27,9 (fourchette = 9,3-46,5). Au total, 30,3% des dispensateurs de soins avaient été formés aux soins aux victimes de viols. Plus des trois quarts des hôpitaux régionaux (76,9%) disposaient d'une salle d'examen réservée aux soins aux victimes de viols. L'analyse de régression multiple des facteurs associés à des soins cliniques de meilleure qualité a mis en évidence que les praticiens concernés avaient plus de 40 ans (estimation du paramètre = 2,4 ; intervalle de confiance (IC) à 95% = 0,7 – 5), qu'ils avaient déjà soigné un plus grand nombre de victimes de viols (estimation du paramètre = 0,02 ; IC à 95% = 0,001 – 0,03), qu'ils travaillaient dans un établissement utilisant un protocole de prise en charge clinique pour soigner les victimes de viols (estimation du paramètre = 2 ; IC à 95% = 0,12 – 3,94), qu'ils travaillaient depuis moins longtemps dans l'établissement (estimation du paramètre = - 0,2 ; IC à 95% = - 0,3 – - 0,04) et qu'ils considéraient le viol comme un problème médical grave (estimation du paramètre = 2,8 ; IC à 95% = 1,9 – 3,8).
CONCLUSION: Les services pour victimes de viols en Afrique du Sud présentent de nombreuses faiblesses. D'après nos observations, il est possible d'améliorer les soins en diffusant des directives relatives à la prise en charge clinique et en veillant à ce que les soins apportés aux victimes de viols soient dispensés par des personnels motivés, spécialement désignés pour les soins aux victimes de viols.

Mots clés: Viol/rééducation & réadaptation; Services santé; Qualité soins; Attitude du personnel soignant; Médecine légale; Enquête système de santé; Etude section efficace; Afrique du Sud (source: MeSH, INSERM).

RESUMEN

OBJETIVO: Centrándonos en el sector de la salud pública en Sudáfrica, decidimos investigar cuáles eran los centros que ofrecían los mejores servicios para las supervivientes de violaciones, quiénes proporcionaban tales servicios, qué actitud tenían los dispensadores de salud ante las mujeres que habían sido violadas, y si la prestación de asistencia para las víctimas tropezaba con algún tipo de problemas.
MÉTODOS: Se llevó a cabo un estudio transversal de establecimientos. En cada una de las nueve provincias de Sudáfrica se muestrearon aleatoriamente dos hospitales de distrito, un hospital regional y un hospital terciario (cuando ello fue posible). En cada hospital, personal superior identificó a dos médicos y dos enfermeras que atendían regularmente a mujeres que habían sido violadas. Se entrevistó a esos médicos y enfermeras mediante un cuestionario que contenía preguntas abiertas y cerradas, de modo que en total aportaron información 124 dispensadores de salud de 31 hospitales. Además, en cada hospital se rellenaba una lista de verificación que indicaba los servicios de que se disponía para las supervivientes de violaciones.
RESULTADOS: El 32,6% de los trabajadores de salud de los hospitales no consideraban que la violación fuese un trastorno médico grave. La media de supervivientes de violaciones atendidas en los seis meses precedentes en cada hospital fue de 27,9 (intervalo = 9,3–46,5). El 30,3% de los proveedores habían recibido formación para atender a víctimas de violaciones. Más de las tres cuartas partes de los hospitales regionales (76,9%) disponían de una sala privada de exploración reservada para esas mujeres. Un análisis de regresión múltiple de las características de los profesionales asociadas a una mejor calidad de la atención clínica puso de manifiesto los siguientes factores relevantes: edad superior a 40 años (estimación del parámetro = 2,4; intervalo de confianza (IC) del 95% = 0,7 - 5), haber atendido antes a un mayor número de víctimas de violaciones (estimación del parámetro = 0,02; IC95% = 0,001 - 0,03), trabajo en un servicio que disponía de un protocolo de manejo clínico de las supervivientes de violaciones (estimación del parámetro = 2; IC95% = 0,12 - 3,94), haber trabajado menos tiempo en el servicio (estimación del parámetro = -0,2; IC95% = -0,3 a -0,04), y el hecho de considerar la violación como un problema médico grave (estimación del parámetro = 2,8; IC95% = 1,9 - 3,8).
CONCLUSIÓN: Los servicios para supervivientes de violaciones de Sudáfrica presentan muchas deficiencias. Nuestros resultados indican que es posible mejorar la atención si se difunden directrices de manejo clínico y se asegura que la asistencia corra a cargo de proveedores motivados específicamente designados para atender a las víctimas.

Palabras clave: Violación/rehabilitación; Servicios de salud; Calidad de la atención de salud; Actitud del personal de salud; Medicina legal; Encuestas de atención de la salud; Estudios transversales; Sudáfrica (fuente: DeCS, BIREME).

Introduction

Rape is an important public health and human rights concern. Its consequences include unwanted pregnancy, unsafe abortion, genital fistulae, pelvic inflammatory disease, sexually transmitted infections such as HIV/AIDS, depression, post-traumatic stress disorder, suicidal behaviour and being socially ostracized (1). These problems are starkly visible in South Africa, which has the highest worldwide annual prevalence of rape reported to the police. Despite documented under-reporting (2), data indicate that in 2002–03 there were 52 425 cases of rape (115 per 100 000 population) reported to police in South Africa compared to, for example, 31.8 per 100 000 in the United States or 8.5 per 100 000 in Brazil (3, 4). In South Africa, we have had reports that patients present to health facilities but do not report the rape to the police. No studies have systematically collected data on how many patients do this.

A key challenge for public health services is to provide adequate gender-sensitive health care for rape survivors within the constraints imposed by locally available resources. In South Africa and other developing countries concerns have been voiced as to how rape survivors are cared for (2, 5). Research on health services for rape survivors can be useful in enhancing the visibility of such problems and improving the quality of care.

Until 1999 in South Africa designated doctors (known as district surgeons) were contracted by the State to deliver care for rape survivors. The services provided by district surgeons were riddled with problems and had been criticized by women's health advocates for many years. An investigation in 1997 by Human Rights Watch found that there was little incentive for district surgeons to "do a good job" (6) . The investigation also found that the system was "deeply flawed with problems of inaccessibility, prejudice and lack of training at all levels" (6). There were often long waits for services, with 1 in 5 patients waiting longer than 5 hours to see a district surgeon after reporting the rape to the police (7).

In 1999 in an effort to improve services and move towards more integrated care, and in keeping with a primary health care approach (8), district surgeons were abolished. Currently, any doctor in public or private practice can provide health services for a rape survivor. However, an exploratory study of services in one province found that changes to services had been made without taking account of the necessity for formal training or for the practitioner to provide evidence of competence. Many doctors were thus ill equipped and reluctant to conduct examinations of rape survivors (9).

It is against this background that an investigation into the quality of services for rape survivors provided by the public health sector in South Africa was undertaken for the Department of Health. Our study aimed to describe aspects of service quality, determine where the best services were provided (whether in tertiary, regional or district hospitals), and to determine which factors influenced the quality of services.

Ethical approval for the study was obtained from the University of Pretoria ethics committee, and the provincial departments of health allowed us to have access to the hospitals.

Methods

Sampling

A cross-sectional study of facilities in all nine provinces of South Africa was undertaken. Two district hospitals, a regional hospital and a tertiary hospital (in provinces with one or more tertiary hospitals) were randomly sampled in all provinces (n = 31) with a probability proportional to the stratum size in each province. The sampling frame consisted of 155 hospitals.

Questionnaires and checklists

At each hospital, the medical superintendent, head of nursing or both were approached and asked to identify for interview two doctors and two nurses who examined or assisted in caring for patients who presented after being raped. In some hospitals there were designated providers who would examine rape survivors while in others any doctor working in casualty was expected to do so.

Data collection instruments (interview questionnaires and facility checklists) were pretested and piloted. Pretesting was carried out at a facility that had not been sampled in Gauteng province. The purpose of the pretest was to ensure that questions were understandable and had face validity. Substantial changes were made to both the questionnaire and the facility checklist after pretesting. A pilot was carried out in one of the sampled districts in Western Cape province. Minor revisions were made to the questionnaire, and the data were included in the final analysis.

Service providers were interviewed face-to-face using a standardized questionnaire that consisted of both open-ended and closed questions. The questionnaire contained five sections: information on the demographic characteristics of providers, what types of services were available for rape survivors, whether protocols for caring for rape survivors were available at the facility, whether the practitioner had undergone training in how to care for rape survivors, and the practitioner's attitudes towards rape and women who have been raped. Since the records of the care received by rape survivors were not comparable between provinces (in some facilities record keeping was done systematically while in others it was chaotic and not accessible) we relied on the practitioners' own reports.

We also asked practitioners to estimate how many rape survivors they had seen during the past six months, whether there had been any problems with the sexual assault evidence collection kit or in sending clothing away for forensic testing, whether they had given evidence in court, whether they had made referrals for psychological support and whether they thought rape was a serious medical problem.

We also gathered information on whether the practitioner raised the possibility of HIV or other sexually transmitted infections with patients after they had been raped, whether they offered an HIV test and pretest and post-test counselling, what advice they gave about post-exposure prophylaxis and whether they prescribed medication for prophylaxis, whether they assessed the patient's risk of pregnancy or offered pregnancy testing, whether they prescribed emergency contraceptives or abortion counselling if indicated, and whether they offered treatment for sexually transmitted infections.

The treatments that practitioners prescribed to presumptively treat sexually transmitted infections were recorded verbatim and then coded after the interview: an answer corresponding to the national syndromic management policy (in which three drugs are prescribed) was awarded 2 points; if two correct drugs were prescribed or a practitioner responded that he or she referred patients for treatment the response was awarded 1 point. Responses to these items were used to develop a scale that measured the quality of clinical care. In addition, fieldworkers completed a checklist at each hospital noting the presence or absence of equipment and medicines and the structural quality of the facilities.

Quality of care

Although there are many factors that affect the quality of services, only the health-service environment and quality of clinical care were measured at the individual practitioner level. Self-reported measures were used at the individual practitioner level. A quality of care composite score consisting of 11 items was developed for the purposes of further data analysis and interpretation (Table 1). This composite score assessed indicators of preventive strategies for sexually transmitted infections and prevention of pregnancy, counselling and the quality of forensic examinations.

Data analysis

Data were analysed with Stata statistical software version 6.0. Using the survey analysis module of the software, the analysis took into account the design of the survey, with respondents clustered into facilities, and the unequal selection probabilities of facilities. Associations between categorical variables were investigated using the Rao–Scott adjustment to the Pearson c² statistic, (as implemented in the svytab command in Stata) (10). A stepwise multiple regression model was built, with backwards elimination, to determine the factors associated with a higher score for quality of care. Candidate variables, in addition to those presented in Table 3, included the level of hospital, staff grade, sex of practitioner and type of training undergone. Open-ended questions were transcribed and analysed thematically.

Findings

Altogether, 31 facilities were sampled. In these facilities, 124 staff were interviewed out of an intended 128 (96.9%). Half of those interviewed were doctors (50.8%) and 64.5% were women. Nearly three-quarters (73.6%) of the women were nurses; 90.8% of the men were doctors (c² = 48.78; F_{1, 13}; P = 0.0001).

We investigated how services were delivered at the different facility levels including tertiary (level 1), regional (level 2) and district (level 3). Results are summarized in Table 2. Doctors conducted most of the medico–legal examinations, with 4.7% of all providers reporting that nurses conducted examinations of rape survivors. Nurses also frequently assisted in examinations conducted by doctors. In an open-ended question asking providers about their role in the examination, providers revealed that nurses sometimes did "superficial" examinations of patients before calling the doctor (Box 1).

Nearly one-third of practitioners working in hospitals (32.6%) did not consider rape to be a serious medical condition. There was a statistically significant difference in this finding among the level of facilities: 12.3% of providers at regional hospitals did not consider rape to be serious compared with 30% at tertiary hospitals and 32.6% at district hospitals (c² = 5.34; F_{1.37, 17.87}; P = 0.04). Altogether, 38% of men and 29.2% of women thought that rape was not a serious medical condition (P = 0.47). If practitioners thought that rape was a serious medical condition it was most often because of the potential health consequences (Box 1). Some providers gave a qualified response, i.e. rape was serious if the patient was a child or there were injuries. Several providers mentioned that women are raped when they are drunk and that women cannot always be believed about rape (data not shown).

The mean number of rape survivors seen in the previous six months was 27.9 (95% confidence interval = 9.3–46.5). Table 2 shows that nearly two-thirds (64.95%) of practitioners had seen fewer than 20 rape survivors in that time but 21.4% had seen 40 or more cases. The frequency of caring for rape survivors varied among facility levels. At district hospitals, 68.4% of practitioners had seen fewer than 20 cases compared with 57.5% of those in tertiary hospitals and 51.2% in regional hospitals.

Few practitioners (14.7%) had ever sent clothing for forensic analysis. One of the possible barriers to this was the lack of emergency clothing available for rape survivors at the facilities. None of the district hospitals had any clothing available, but 32.3% of regional hospitals and 50% of tertiary hospitals did (P = 0.002).

There was a short poorly-disseminated national protocol for the care and management of rape survivors which some provinces had adapted. However, more than half (59.1%) of the providers reported that there was no protocol for the care of rape survivors where they worked (Table 2).

Less than one-third (30.3%) of providers had ever received training in caring for rape survivors. Nearly half of those who had had received their training as undergraduates. Most training covered medical treatment: 93.2% of those who had been trained said that training covered medical treatment and 88.6% had received training in collecting forensic specimens. Only 34.8% of those trained said that gender issues had been discussed; 50% said that psychosocial aspects had been addressed.

Whether a private room with four walls and a door for examining rape survivors was available varied according to the level of facility (Table 2). None of the tertiary (level 1) facilities had a private examination room. HIV tests were available in the examination room in 20% of tertiary hospitals, 53.9% of regional hospitals and 60.7% of district hospitals.

Providers said they routinely discussed the risks of HIV, sexually transmitted infections and pregnancy with patients, and there was no difference among the different levels of hospitals. When the results from facilities were aggregated we found that all facilities were significantly less likely to offer pretest counselling for HIV than they were to offer an HIV test (c² = 59.07; F_{1, 13}; P < 0.0001). Although the study took place during a period when government policy changed to allow for the provision of post-exposure HIV prophylaxis, only 19.7% of practitioners provided this.

Altogether 70.6% of practitioners reported asking patients about contraceptive use; 71.2% offered a pregnancy test; and 84% offered emergency contraception. When patients present at a facility five days or more after being raped, abortion counselling may be required. One-quarter of practitioners offered abortion counselling. This was much more common in regional hospitals than in hospitals at other levels (P = 0.007).

Sexually transmitted infections were treated by 93.2% of practitioners. However, only 36.9% named the correct drugs for this, and there was no difference in this result among facilities (P = 0.45).

Less than half (48.8%) of practitioners reported that they referred patients for counselling after rape. There was a statistically significant difference between level of facility: providers in 72.5% of tertiary facilities referred patients for counselling compared with those in 71.5% of regional facilities and 35.1% of district facilities (c² = 8.45; F_{1.88, 24.42}; P = 0.0015). A variety of reasons were given including that providers did not know to whom they could refer patients or that it was not their role to refer for counselling. A few providers mentioned that they had never thought of referring patients.

Nearly half (46.2%) of the providers reported having problems with the sexual assault evidence collection kits. The kits were mostly kept by the police and brought to the health facility by the patient if she went first to the police. Providers reported that kits were sometimes incomplete, had already been used, were unavailable or that the police forgot to bring them. Doctors frequently mentioned that a new sexual assault evidence collection kit had been introduced and they had not been trained to use it, and they often did not know how to use the different components.

Only 15.2% of facilities were found to have a lockable cupboard for storing evidence (Table 2). Thus completed kits were often left on a nurse's desk for collection by the police; this is in clear breach of requirements to protect the chain of evidence from possible tampering.

Practitioners in hospitals were asked whether they had given evidence in court during the previous year. Few (11.5%) had done so. None of the nurses, including the forensically trained nurses, had given evidence in court.

Multiple regression analysis of factors associated with providing a higher quality of clinical care suggested that older staff who had cared for a higher number of rape survivors, who worked in a facility with a protocol for treating rape survivors and who perceived rape to be a serious medical problem provided better care (Table 3). The best management was associated with the highest quartile of the caseload (providers who had seen more than 40 rape survivors in the previous 6 months). Those who had been working in the facility for a longer time scored lower on measures of quality of care, possibly because of burn out.

Discussion

This study has highlighted many weaknesses in the facilities for, and the care of, rape survivors. No definitive pattern of better management emerged when the different levels of hospitals were compared. However the data suggested that providers at regional facilities were more likely to have gender-sensitive attitudes, such as thinking that rape was a serious medical condition, and they were more likely to offer abortion counselling, when indicated. Within the constraints of the current service, i.e. where few staff have been trained in caring for rape survivors, the best care was provided by older staff who had had more experience caring for rape survivors and who worked in a facility with guidelines on caring for such patients, and who believed rape to be a serious problem but who had not worked at the facility for too long. This suggests that it is preferable to provide a service with a smaller number of dedicated providers. The integration into clinics of care for rape survivors may be undesirable because there are many clinics and so, inevitably, practitioners at these clinics will for the most part care for few rape survivors.

In a study conducted by Ledray & Simmelink in 1997 in the United States trained providers were found to collect better evidence and maintain the chain of evidence better when compared with other medical professionals who used a standardized rape examination kit without prior training (11). In South Africa training, whether given before or after qualification, was not associated with providing better quality care. This could reflect the brevity of the training; the content, which focuses primarily on the medico–legal aspects of rape; or the lack of interest among service providers. Attitudes are hard to change, but these findings suggest that training on caring for rape survivors needs to emphasize its social context and the importance of the medical and social management of women. Furthermore, the small proportion of practitioners who have had any training in caring for rape survivors needs to be addressed.

While there are some initiatives under way in South Africa to introduce the concept of nurse examiners or forensic nurses who care for survivors of sexual assaults, this is not widespread. Findings from this study suggest that few of the nurses interviewed had taken part of any of these initiatives.

The lack of training may explain the poor practice regarding the preservation and maintenance of the chain of evidence. Nurses' preliminary examinations could potentially result in the loss of physical evidence. Other practices, such as keeping completed sexual assault examination kits in public places, may be another consequence, and this oversight would render the findings of the analysis inadmissible in court (12). The use of private rooms designated specifically for the examination of patients assists in building a sense of security, privacy and confidentiality for the patient, as well as in maintaining the chain of evidence (13).

Providers are often reluctant to care for patients who have been raped because they may be required to give evidence in court (9, 14, 15). Anecdotal evidence suggests that this is a barrier to training nurses as sexual assault examiners because they believe that their expert testimony may not carry the same weight as that of doctors (K. Müller, personal communication, 2004). However, our findings show that providers rarely attend court: in fact, few had ever done so. We were unable to determine why providers had not given evidence. It may be a reflection of the small proportion of rape cases that actually go to trial but we cannot exclude the possibility that written evidence was good enough to make oral evidence unnecessary.

The interaction between police and health services needs to be improved. The current system, which requires the police to bring the sexual assault evidence collection kit to the hospital, has an impact on the extent to which health services are client centred. This system requires patients to present first at the police station or to wait for the police to bring the kit to the hospital. If evidence collection kits were kept at health facilities, the waiting time for patients might be reduced. Keeping the kits at the police station also furthers the dominant belief that rape is predominantly a criminal justice concern, rather than a health concern. Many practitioners viewed their responsibility as the collection of evidence rather than the care of the patient. This is particularly evident in the lack of referrals for psychological support: many practitioners did not see this as an integral part of their role.

The management of sexually transmitted infections or unwanted pregnancy is not specific to the care of women who have been raped. The poor management identified in this study indicates that the quality of care for reproductive health in general may not be optimal and should be addressed. It is important that processes to introduce post-exposure HIV prophylaxis after rape also embrace improvements in preventing and treating sexually transmitted infections and unwanted pregnancy.

This study has its limitations. For example, staff were not randomly selected for interview within facilities, however by interviewing nominated staff we hoped to capture those most involved in caring for rape survivors. Additionally, information on training could have been subject to recall bias, and all care (including the number of rape survivors seen) was self-reported and could not be verified.

Conclusion

There are many gaps in the care available to South Africa's rape survivors. This research has been important in rendering these gaps visible and can be used by policy-makers to identify the key interventions needed to improve care. Given the resource constraints of a middle-income country, it is important to be able to demonstrate the value of clinical guidelines and the relative merits of a service provided by designated providers versus a generalized service.

Ultimately, rape is a gender issue. It is interesting to see that staff who understood the meaning of rape in terms of women's lives provided better services. The methods developed for this study have been adopted by WHO for use in other developing countries (16). This research has the potential to substantially advance our understanding of services for rape survivors and assist in international efforts to improve them.

Acknowledgements

Josephine Malala, Zanele Silo, Nwabisa Jama and Tanya Jacobs helped with the data collection and Dr Jonathan Levin commented on the statistical analysis. We would like to thank all the doctors and nurses who agreed to be interviewed for this study.

Funding: This work was funded by the Rockefeller Foundation through their support for the South African Gender-Based Violence and Health Initiative (SAGBVHI).

Competing interests: none declared.

References

1. Jewkes R, Sen G, Garcia-Moreno C. Sexual violence. In: Krug EG, Dahlberg LL, Mercy JA, Zwi AB, Lozano R, editors. World report on violence and health. Geneva: World Health Organization; 2002. p. 147–74.        

2. Jewkes R, Abrahams N. The epidemiology of rape and sexual coercion in South Africa: an overview. Social Science and Medicine. 2002;55:1231-44.        

3. South African Police Services. Rape in the RSA for the financial years 1994/1995 to 2003/2004. Pretoria: South African Police Services, Crime Information Analysis Center; 2004.          Available from: http://www.saps.gov.za/statistics/reports/crimestats/2004/_pdf
/crimes/rsa_totals03_04_new.pdf

4. Interpol. International crime statistics: 2004. Lyon: Interpol; 2004.        

5. Kim JC, Martin LJ, Denny L. Rape and HIV post-exposure prophylaxis: addressing the dual epidemics in South Africa. Reproductive Health Matters 2003;11:101-12.        

6. Human Rights Watch. South African violence against women and the medico–legal system. New York: Human Rights Watch; 1997.        

7. Swart L, Gilchrist A, Butchard A, Seedat M, Martin L. Rape surveillance through district surgeons' offices in Johannesburg, 1996-1998: evaluation and prevention implications. Pretoria: Institute of Social and Health Sciences, University of South Africa; 1999.        

8. McIntyre D, Klugman B. The human face of decentralization and integration of health services: experiences from South Africa. Reproductive Health Matters 2003;11:108-19        

9. Suffla S, Seedat M, Nascimento A. Evaluation of medico-legal services in Gauteng: implications for the development of best practices in the after-care of rape survivors. Pretoria: Institute for Social and Health Sciences and Centre for Peace Action, University of South Africa; 2001.        

10. Rao JNK, Scott AJ. The analysis of categorical data from complex sample surveys: chi-squared tests for goodness of fit and independence in two-way tables. Journal of the American Statistical Association 1981;76:221-30.        

11. Ledray LE, Simmelink K. Efficacy of SANE evidence collection: a Minnesota study. Journal of Emergency Nursing 1997;23:75-7.        

12. World Health Organization.Guidelines for the medico–legal care of victims of sexual violence. Geneva: WHO; 2003.        

13. World Health Organization. Guidelines for medico-legal care of victims of sexual violence. Geneva: WHO; 2003.        

14. Department of Health and South African Gender-Based Violence and Health Initiative. Developing an appropriate health sector response to gender-based violence: report of a workshop jointly hosted by the National Department of Health and the South African Gender-Based Violence and Health Initiative. Pretoria: DOH, SAGBVHI; 2001.        

15. Green WM, Panacek EA. Sexual assault forensic examinations in evolution. Journal of Emergency Medicine 2003;25:97-9.        

16. World Health Organization. The sexual violence research initiative. Available from: http://www.who.int/svri/en/        

(Submitted: 20 July 2004 – Final revised version received: 7 January 2005 – Accepted: 13 January 2005)

1 Correspondence should be sent to Dr Christofides at this address (email: nicola.chris@mrc.ac.za).