Forensic Nurse Examiners versus Doctors for the Forensic Examination of Rape and Sexual Assault Complainants: A Systematic Review

Abstract

BACKGROUND

Within the UK, the complainants of rape and sexual assault are typically referred to regional sexual assault referral centres (SARCs) where their medical and psychological needs are addressed and, if they consent, a forensic medical examination will be conducted, usually by a forensic physician. In the USA, this service is typically nurse-led.

OBJECTIVES

To compare the reliability and efficacy of Sexual Assault Nurse Examiners (SANEs)/Forensic Nurse Examiners (FNEs) with that of non-SANE health professionals in the conduct of the forensic medical examination and the collection of forensic evidence (rape kit) from the complainants of rape and sexual assault. The following outcomes are used to quantify the efficacy of the SANEs: complainant quality of life, conviction and prosecution rates, complainant mortality within 30 days, time from complaint to examination, provision of STI, pregnancy and HIV prophylaxis, collection and documentation of rape kits and forensic examination, number of rape kits admissible as evidence, and the average cost per case.

SEARCH STRATEGY

The following databases were searched: The Cochrane Library, MEDLINE, EMBASE, AMED, CINAHL, PsychInfo, BNI, Health Business Elite, HMIC, Social Policy and Practice, Google Scholar, and the Scientific Citation Index. Relevant studies were selected by two independent reviewers and no restrictions were placed on either the year or language of publication.

SELECTION CRITERIA

This review included studies comparing outcomes for complainants of rape and/or sexual assault who were treated by a SANE, with those treated by a non-SANE health professional, irrespective of the study design and the age of the complainants.

DATA COLLECTION AND ANALYSIS

Two reviewers were involved in the data collection and analysis. Risk ratios (RR) or mean differences (MD) with 95% confidence intervals (95% CI) were calculated with both the random-effects and fixed-effects model using RevMan 5.1 software. Where differences were noted between the results, both models have been reported. Where no significant differences have been found, only the results from the random-effects model are reported (data from both models can be found in Appendix 1).

RESULTS

A total of eight studies were included in the systematic review, six of which were included in the meta-analysis. This provided an overall sample size of 2700 complainants; 1223 complainants were examined by a SANE (SANE group) and 1477 were examined by a non-SANE health professional (non-SANE group). No data were available on complainant quality of life. Two studies compared the conviction and prosecution rates, with no significant differences found (relative risk (RR) 1.00, 95% confidence interval (CI) 0.64 to 1.55 and RR 1.04, 95% CI 0.73 to 1.48 respectively). Significantly more rape kits in the SANE group were admissible as evidence in court (RR 1.20, 95% CI 1.06 to 1.35). No data were reported for 30-day mortality or time from complaint to examination.

In terms of clinical care, complainants in the SANE group were significantly more likely to have received STI and pregnancy prophylaxis than those in the non-SANE group (RR 1.07, 95% CI 1.01 to 1.13 and RR 1.32, 95% CI 1.19 to 1.46 respectively). No significant differences were found regarding the provision of HIV prophylaxis (RR 1.29, 95% CI 0.87 to 1.89). Using a fixed-effects model, complainants in the SANE group were significantly more likely both to have a forensic examination (rape kit) and to have it documented (RR 3.94, 95% CI 3.21 to 4.84 and RR 3.21, 95% CI 2.71 to 3.80 respectively). However, the results were not significant with a random-effects meta-analysis (RR 2.79, 95% CI 0.21 to 36.38 and RR 2.28, 95% CI 0.65 to 8.01).

In terms of cost, the SANEs were found, on average, to be £68 cheaper per case than their physician counterparts. Confidence interval data were not available for this outcome and it is not clear if this difference is significant.

AUTHORS' CONCLUSIONS

While there does not appear to be any benefit gained in terms of prosecution and conviction by substituting forensic doctors with forensic nurse examiners (FNEs), the FNEs do seem to be statistically significantly better in the provision of clinical care and are able to provide a cheaper service than that led by physicians. However, due to the limited data available to this review, it should be borne in mind that the evidence base for these conclusions is very weak, and, without further research, should not necessarily be used to form the basis for any significant services changes.

Campbell Systematic Reviews

2014:5

First published: 01 May, 2014 Search executed February, 2012

Forensic Nurse Examiners versus Doctors for the Forensic Examination of Rape and Sexual Assault Complainants: A Systematic Review

Clare Toon & Kurinchi Gurusamy

Colophon

Title Forensic Nurse Examiners versus Doctors for the Forensic Examination of

Rape and Sexual Assault Complainants: A Systematic Review Institution The Campbell Collaboration

Authors Toon, Clare

Gurusamy, Kurinchi DOI 10.4073/csr.2014.5 No. of pages 56

Last updated 25 January, 2013 Citation Toon C, Gurusamy K. Forensic Nurse Examiners Versus Doctors for the

Forensic Examination of Rape and Sexual Assault Complainants: A

Systematic Review

Campbell Systematic Reviews 2014:5

DOI: 10.4073/csr.2014.5 ISSN 1891-1803

Copyright Toon & Gurusamy

This is an open-access article distributed under the terms of the Creative

Commons Attribution License, which permits unrestricted use, distribution,

and reproduction in any medium, provided the original author and source are

credited.

Contributions Clare Toon conceived the review question, developed the review, performed

the writing of the review, made an intellectual contribution and approved the

final version prior to submission.

Kurinchi Gurusamy made an intelectual contribution, advised on the review

and approved the final version prior to submission. Support/funding This review was conducted as part of a final MSc project and received no

financial support Potential conflicts

of interest

Clare D Toon

Clinical Effectiveness Officer

Public Health

WSCC County Hall Campus

Grange Block (first floor), West Street

Chichester, West Sussex

E-mail: [email protected]

The authors have no vested interest in the outcomes of this review, nor any

incentive to represent findings in a biased manner. Corresponding

author

Campbell Systematic Reviews

Editors-in-Chief Arild Bjrndal, The Centre for Child and Adolescent Mental Health, Eastern

and Southern Norway & University of Oslo, Norway

Julia Littell, Bryn Mawr College, USA

Editors

Crime and Justice David B. Wilson, George Mason University, USA Education Sandra Wilson, Vanderbilt University, USA International

Development

Peter Tugwell, University of Ottawa, Canada

Howard White, 3ie, India

Methods Therese Pigott, Loyola University, USA

Ian Shemilt, University of Cambridge, UK

The Campbell Collaboration (C2) was founded on the principle that

systematic reviews on the effects of interventions will inform and help

improve policy and services. C2 offers editorial and methodological support to

review authors throughout the process of producing a systematic review. A

number of C2's editors, librarians, methodologists and external peer-

reviewers contribute.

The Campbell Collaboration

P.O. Box 7004 St. Olavs plass

0130 Oslo, Norway

http://www.campbellcollaboration.org/

Web End =www.campbellcollaboration.org

Birte Snilstveit, 3ie, UK

Hugh Waddington, 3ie, UK

Social Welfare Nick Huband, Institute of Mental Health, University of Nottingham, UK

Geraldine Macdonald, Queens University, UK & Cochrane Developmental,

Psychosocial and Learning Problems Group Methods Therese Pigott, Loyola University, USA

Emily Tanner-Smith, Vanderbilt University, USA Managing Editor Karianne Thune Hammerstrm, The Campbell Collaboration

Co-Chairs

Crime and Justice David B. Wilson, George Mason University, USA

Martin Killias, University of Zurich, Switzerland Education Paul Connolly, Queen's University, UK

Gary W. Ritter, University of Arkansas, USA Social Welfare Aron Shlonsky, University of Toronto, Canada

Jane Barlow, University of Warwick, UK International

Development

Table of contents

TABLE OF CONTENTS 3

ABSTRACT 5 Background 5 Objectives 5 Search Strategy 5 Selection Criteria 6 Data collection and Analysis 6 Results 6 Authors Conclusions 7

1 BACKGROUND 8 1.1 Description of the condition 8 1.2 Description of the intervention 9 1.3 How the intervention might work 9 1.4 Why it is important to do this review 10

2 OBJECTIVES 11

3 METHODS 12 3.1 Criteria for considering studies in this review 12 3.2 Search methods for identification of studies 13 3.3 Data collection and analysis 14

4 MAIN RESULTS 17 4.1 Description of studies 18 4.2 Risk of bias in included studies 19 4.3 Effects of intervention 21 4.4 Statistical variations 25 4.5 Heterogeneity 25 4.6 Publication bias 26 4.7 Summary of findings 26

5 DISCUSSION 29

6 CONCLUSIONS 35 6.1 Implications for practice 35 6.2 Implications for research 35

3 The Campbell Collaboration | www.campbellcollaboration.org

7 REFERENCES 37

8 REFERENCES FOR ELIGIBLE STUDIES 40

9 REFERENCES FOR EXCLUDED STUDIES 41

10 TABLES 44

11 SEARCH STRATEGIES 59

4 The Campbell Collaboration | www.campbellcollaboration.org

Abstract

BACKGROUND

Within the UK, the complainants of rape and sexual assault are typically referred to regional sexual assault referral centres (SARCs) where their medical and psychological needs are addressed and, if they consent, a forensic medical examination will be conducted, usually by a forensic physician. In the USA, this service is typically nurse-led.

OBJECTIVES

To compare the reliability and efficacy of Sexual Assault Nurse Examiners (SANEs)/Forensic Nurse Examiners (FNEs) with that of non-SANE health professionals in the conduct of the forensic medical examination and the collection of forensic evidence (rape kit) from the complainants of rape and sexual assault. The following outcomes are used to quantify the efficacy of the SANEs: complainant quality of life, conviction and prosecution rates, complainant mortality within 30 days, time from complaint to examination, provision of STI, pregnancy and HIV prophylaxis, collection and documentation of rape kits and forensic examination, number of rape kits admissible as evidence, and the average cost per case.

SEARCH STRATEGY

The following databases were searched: The Cochrane Library, MEDLINE, EMBASE, AMED, CINAHL, PsychInfo, BNI, Health Business Elite, HMIC, Social Policy and Practice, Google Scholar, and the Scientific Citation Index. Relevant studies were selected by two independent reviewers and no restrictions were placed on either the year or language of publication.

SELECTION CRITERIA

This review included studies comparing outcomes for complainants of rape and/or sexual assault who were treated by a SANE, with those treated by a non-SANE health professional, irrespective of the study design and the age of the complainants.

5 The Campbell Collaboration | www.campbellcollaboration.org

DATA COLLECTION AND ANALYSIS

Two reviewers were involved in the data collection and analysis. Risk ratios (RR) or mean differences (MD) with 95% confidence intervals (95% CI) were calculated with both the random-effects and fixed-effects model using RevMan 5.1 software. Where differences were noted between the results, both models have been reported. Where no significant differences have been found, only the results from the random-effects model are reported (data from both models can be found in Appendix 1).

RESULTS

A total of eight studies were included in the systematic review, six of which were included in the meta-analysis. This provided an overall sample size of 2700 complainants; 1223 complainants were examined by a SANE (SANE group) and 1477 were examined by a non-SANE health professional (non-SANE group). No data were available on complainant quality of life. Two studies compared the conviction and prosecution rates, with no significant differences found (relative risk (RR) 1.00, 95% confidence interval (CI) 0.64 to 1.55 and RR 1.04, 95% CI 0.73 to 1.48 respectively). Significantly more rape kits in the SANE group were admissible as evidence in court (RR 1.20, 95% CI 1.06 to 1.35). No data were reported for 30-day mortality or time from complaint to examination.

In terms of clinical care, complainants in the SANE group were significantly more likely to have received STI and pregnancy prophylaxis than those in the non-SANE group (RR 1.07, 95% CI 1.01 to 1.13 and RR 1.32, 95% CI 1.19 to 1.46 respectively). No significant differences were found regarding the provision of HIV prophylaxis (RR 1.29, 95% CI 0.87 to 1.89). Using a fixed-effects model, complainants in the SANE group were significantly more likely both to have a forensic examination (rape kit) and to have it documented (RR 3.94, 95% CI 3.21 to 4.84 and RR 3.21, 95% CI 2.71 to 3.80 respectively). However, the results were not significant with a random-effects meta-analysis (RR 2.79, 95% CI 0.21 to 36.38 and RR 2.28, 95% CI 0.65 to 8.01).

In terms of cost, the SANEs were found, on average, to be 68 cheaper per case than their physician counterparts. Confidence interval data were not available for this outcome and it is not clear if this difference is significant.

AUTHORS CONCLUSIONS

While there does not appear to be any benefit gained in terms of prosecution and conviction by substituting forensic doctors with forensic nurse examiners (FNEs),

6 The Campbell Collaboration | www.campbellcollaboration.org

the FNEs do seem to be statistically significantly better in the provision of clinical care and are able to provide a cheaper service than that led by physicians. However, due to the limited data available to this review, it should be borne in mind that the evidence base for these conclusions is very weak, and, without further research, should not necessarily be used to form the basis for any significant services changes.

7 The Campbell Collaboration | www.campbellcollaboration.org

1 Background

1.1 DESCRIPTION OF THE CONDITION

The Sexual Offences Act defines sexual assault as Intentional contact, of a sexual nature, with another persons body without their consent or knowledge (Home Office, 2003). More specifically, the act defines rape as Intentional penetration of the vagina, anus, or mouth with a penis, without consent. In 2010/11 there were a total of 54,982 sexual offences (including both rape and sexual assault) reported by the police in England and Wales (Chaplin, 2011). In addition to this, a further 17,727 sexual crimes were reported against children (National Society for the Prevention of Cruelty to Children, 2012). While the reported incidence of rape and sexual assault is already relatively high, it has been suggested that around 40% of adult and 31% of paediatric complainants of rape and sexual assault do not disclose their experience(s) to anyone (HM Government, 2007). As such, current reports of the incidence of rape and sexual assault may be well below the true figure.

It has been estimated that each rape costs 73,487 (Dubourg and et al., 2005), and the overall cost of sexual violence to society was estimated at 8.5 billion in 2003/04 alone (HM Government, 2007). When investigating the costs involved in such crime, it is worth noting that the figures quoted above only include the costs to society and not personal costs incurred by the complainants of sexual violence. It is also important to note that neither the long-term emotional, physical or psychological consequences of rape or sexual assault appear to have been considered in this costing exercise.

Current best practice in assisting and treating the complainants of rape and sexual assault within the UK is through the use of a Sexual Assault Referral Centre (SARC) (Home Office et al., 2005). The first of these centres was established in 1986 within St Marys Hospital in Manchester (Home Office et al., 2005). The purpose of any SARC is to offer the complainants of rape and sexual assault the support of crisis workers, forensic examination, appropriate medical care, and psychological support throughout their experience of the medical, forensic and legal aspects of their experience. In addition to improving care of the complainants, this approach to rape and sexual assault is thought to offer an increased standard of forensic evidence collection and, subsequently, increases the potential for the conviction and prosecution of the perpetrators of sexual offences (Home Office et al., 2005).

8 The Campbell Collaboration | www.campbellcollaboration.org

There are currently 30 SARCs in operation across the UK, with a further 15 planned (Stern, 2010) and, in addition to improving care at the point of contact, they allow self-referring complainants access to a forensic examination, should this be their wish. Forensic examinations are typically carried out by female forensic physicians; however, there are three specialist nurses currently carrying out this role, with a further 20 in training (UKAFN, 2012a). Prior to the implementation of the SARCs, any complainant wishing to undergo forensic examination was required to report their assault to the police. Medical and forensic services were typically delivered through busy emergency departments, police stations and, in one case, a hospital mortuary, where care was administered by unspecialised doctors and nursing staff. The implementation of the SARCs meant that any complainant of rape could choose to undergo a forensic examination, without first contacting the police. The forensic evidence from such examinations is stored so that the complainant can, if they wish to, pursue prosecution at a later date.

1.2 DESCRIPTION OF THE INTERVENTION

Forensic Nurse Examiners (FNEs) are fully qualified nurses, who have been trained in the field of forensic evidence gathering, with particular respect to cases of rape and sexual assault in both adults and children (IAFN, 2006). Although there are only three FNEs currently working within the UK, the American/Canadian equivalent of these nurses, the Sexual Assault Nurse Examiners (SANEs), have been established since the late 1970s (Ledray and Simmelink, 1997). Like the SARCs in the UK, SANE programmes in the USA were established in response to issues surrounding the proper care of rape and sexual assault complainants and, more specifically problems surrounding the recruitment and retention of female forensic examiners to deal with these cases (Ledray and Simmelink, 1997, Kelly, 2004). The SANE is typically responsible for conducting forensic examinations, collecting and documenting findings, preparing statements for court when requested to by the police or Criminal Prosecution Service (CPS), and giving evidence in court, where necessary (Kelly, 2004). As noted above, there are currently very few FNEs employed within SARCs in the UK. Unlike their SANE counterparts, the FNEs do not currently attend court and their statements are checked and verified by a medical doctor prior to any case going to court (Inott, 2012).

1.3 HOW THE INTERVENTION MIGHT WORK

The intervention would work by substituting forensic doctors with FNEs as the lead in the forensic and clinical aspects of rape and sexual assault cases. One of the most obvious benefits of this would be a reduction of the costs involved, as retention of specialised forensic doctors is generally more expensive than that of nursing specialists (NHS Employers, 2012, Royal College of Nursing, 2012, Dubourg and et

9 The Campbell Collaboration | www.campbellcollaboration.org

al., 2005, Kelly, 2004). Additionally, nurses are able to spend more time with the complainants, show greater empathy, and are better able to foster a sense of control and empowerment in the complainants of sexual crimes (Campbell et al., 2008, Campbell et al., 2005). Further benefit would be the highly specialised nature of this role, which would allow these nurses to hone their skills and develop their expertise and, potentially, maximise their credibility and acceptability as expert witnesses in a court of law. Also, by employing specialist nurses, it is more likely that all SARCs will be in a position to allow the complainants of rape and sexual assault to choose the gender of their examiner, as required by current national guidelines (Department of Health et al., 2009).

1.4 WHY IT IS IMPORTANT TO DO THIS REVIEW

There are currently no systematic reviews examining the efficacy and reliability of forensic nurses against that of non-SANE health professionals in the care and treatment of the complainants of rape and sexual assault. As such, this review may prove useful in informing the organisation and commissioning of sexual assault services in the longer term.

10 The Campbell Collaboration | www.campbellcollaboration.org

2 Objectives

To compare the reliability and efficacy of Sexual Assault Nurse Examiners (SANEs)/Forensic Nurse Examiners (FNEs; henceforth referred to as SANEs) with that of non-SANE health professionals in the conduct of the forensic medical examination and collection of forensic evidence from complainants of rape and sexual assault. The following outcomes are used to quantify the efficacy of the SANEs: complainant quality of life, conviction and prosecution rates, complainant mortality within 30 days, time from complain to examination, provision of STI, pregnancy and HIV prophylaxis, collection and documentation of rape kits and forensic examination, number of rape kits admissible as evidence, and the average cost per case.

11 The Campbell Collaboration | www.campbellcollaboration.org

3 Methods

3.1 CRITERIA FOR CONSIDERING STUDIES IN THIS REVIEW

Types of studies:

Both experimental and quasi-experimental trial designs were eligible for inclusion in this review. Quasi-experimental designs were restricted to designs that included a control or comparator condition. No restrictions were placed on publication language, publication date or status, or the size of the sample included. Case-studies or case-series were not included as they do not offer the opportunity to compare the two different types of intervention being investigated here.

Types of participants:

All complainants of rape or sexual assault, regardless of their age or gender, or whether or not they reported the incident(s) to the police were considered eligible participants for this review. The intervention group comprised those complainants who had been examined and treated by a Sexual Assault Nurse Examiner (SANE); the control group comprised those who were examined and treated by a non-SANE health professional. Participants were included where data relating to at least one of the outcome measures was available.

Types of intervention:

Forensic medical examination of rape complainants which have been conducted by a SANE, who has been fully trained in the field of forensic evidence gathering, with particular respect to cases of rape and sexual assault in both adults and children.

Types of control:

Forensic medical examination of rape complainants which have been carried out by a non-SANE health professional.

12 The Campbell Collaboration | www.campbellcollaboration.org

Types of outcome measures: Primary outcomes:

1. Complainants quality of life. 2. Proportion of cases resulting in conviction. 3. Proportion of cases resulting in prosecution. 4. Complainant mortality within 30 days.

Secondary outcomes:

1. Time from complaint to examination. 2. Provision of sexually transmitted infection (STI), pregnancy and human immunodeficiency virus (HIV) prophylaxis.

3. Proportion of complainants who had a rape kit collected. This is a collection of evidence, which is used for prosecution purposes.

4. Proportion of complainants who had a forensic examination documented. This is where it has been documented that a forensic examination has been carried out, regardless of whether or not a rape kit has been collected.

5. Number of rape kits admissible as evidence. 6. Cost per case.

Evidential quality was assessed through the analysis of data collected for the proportion of complainants who had a rape kit collected, those who had a forensic examination documented and the number of rape kits which were considered admissible as evidence.

3.2 SEARCH METHODS FOR IDENTIFICATOIN OF STUDIES

Electronic searches:

The following databases were searched: The Cochrane Library, MEDLINE, EMBASE, AMED, CINAHL, PsychInfo, BNI, Health Business Elite, HMIC, Social Policy and Practice, Google Scholar, and the Scientific Citation Index. The search strategies can be found in Appendix 1.

Searching other resources:

The following professional organisations and bodies were contacted in order to identify any additional unpublished reports and data: Rape Crisis (an organisation which exists to promote the needs of women and girls, who have experienced sexual violence), all SARCs registered in the UK, UK Association of Forensic Nurses

13 The Campbell Collaboration | www.campbellcollaboration.org

UKAFN), the International Society of Forensic Nurses (IAFN), both the Canadian and American Society of Forensic Nurses, and the Worshipful Society of Apothecaries of London (an organisation offering training to health professionals and academic examination in the forensic and clinical aspects of sexual assault).

3.3 DATA COLLECTION AND ANALYSIS

Selection of studies:

Two reviewers worked independently to identify studies for inclusion in this review, and to list the reasons for studies which have been excluded. Any differences of opinion were resolved through discussion.

Data extraction and management:

In order to ensure accuracy and reduce the potential for bias within the review, data were extracted independently by two reviewers, using an Excel data extraction form. In the event that a full data set was not available, the individual study authors were contacted with the aim of obtaining this information. In addition to data for the outcomes detailed above, the following data were also extracted: 1. Year of publication. 2. Country. 3. Inclusion and exclusion criteria. 4. Details of the interventions. 5. Total sample size and the number of participants in each treatment group. 6. Gender and mean age of the participants. 7. Outcomes reported. 8. Bias risk assessment.

Any differences of opinion were resolved through discussion.

Assessment of risk of bias in included studies:

Bias can be referred to as prejudice or systematic error. If there is bias present within any of the included studies, this can lead to a misrepresentation of the truth and, consequently can lead to an over- or under-estimation of the true effect of the intervention under investigation. Risk of bias within the included studies was assessed in accordance with the guidelines set down by the Cochrane Collaboration (Higgins and Green, 2011). The following components were assessed:

1. Selection bias this refers to systematic difference in group selection and, consequently, differences in the baseline characteristic between the two intervention groups.

14 The Campbell Collaboration | www.campbellcollaboration.org

2. Blinding of participants and personnel also referred to as performance bias. This refers to systematic differences in the care provided to participants in the two intervention groups. Consequently, the two groups are exposed to different factors.

3. Blinding of outcome assessment also known as detection bias. This refers to systematic differences in how the outcomes are determined.

4. Incomplete outcome data typically known as attrition bias. This refers to systematic differences between the two intervention groups in terms of withdrawals from the study.

5. Selective reporting this refers to systematic differences between reported and unreported findings. It can also be observed when the order of the outcomes is altered.

6. Vested-interest bias this can be observed as study design, analysis, and presentation of results which is biased in favour of the funder and/or researcher.

These elements were individually classified as high, low, or unclear risk of bias, based on the information available from each individual study.

Measures of treatment effect:

For all of the binary outcomes, risk ratios (RR) were calculated, along with 95% confidence intervals (95% CI). None of the continuous outcomes were reported.

Unit of analysis issues:

The individual complainants of rape or sexual assault were treated as the units of analysis.

Dealing with missing data:

No data were missing for the outcomes reported by any of the included studies.

Assessment of heterogeneity:

Statistical heterogeneity was established through the use of the Chi-squared test. As the Chi-squared test is relatively low-powered, a P-value of 0.10 was used to indicate heterogeneity. The magnitude of any heterogeneity was quantified using Higgins I (Higgins and Thompson, 2002). An I below 40% was considered to be unimportant; 30% 60% indicated moderate heterogeneity; 50% - 90% substantial heterogeneity; and an I in excess of 75% indicated considerable heterogeneity (Higgins and Green, 2011).

Assessment of reporting biases:

Not enough studies were included within this analysis to allow the accurate detection of any reporting bias.

15 The Campbell Collaboration | www.campbellcollaboration.org

Data synthesis:

The meta-analysis was conducted using RevMan 5.1, using both a fixed-effects model (Demets, 1987) and a random-effects model (DerSimonian and Laird, 1986). When combining the data, the fixed-effect model assumes that each study is estimating exactly the same quantity, and produces the best estimate of the mean effect. However, the random-effects model makes no such assumption. With this approach it is assumed that that estimated intervention effects follow a distribution across studies. The result produced by the random-effects model is the average of the mean effects in different populations. Where no differences were found between the two models the results from the random-effects model have been reported (data from both models can be found in Appendix 1). Where there were discrepancies affecting the significance of the effect estimates, both models are reported. The Mantel-Haenszel method was used for the binary outcomes and the inverse variance method was used for the continuous outcomes.

Subgroup analysis and investigation of heterogeneity:

The following subgroup analyses were planned:

1. Medical doctors versus other emergency medical personnel as non-SANE health professionals in the control group.

2. UK versus USA programmes. 3. Paediatric versus adult cases.

However, there were too few studies to make these analyses possible. There were also too few studies to allow the conduct of sensitivity analyses to establish the cause of any observed heterogeneity. With regard to heterogeneity, it is worth noting that all data were included in the analyses, regardless of any variations in the provision of care between the studies.

16 The Campbell Collaboration | www.campbellcollaboration.org

4 Main Results

Although randomised controlled trials (RCTs) were eligible, none were identified. While this does leave this review prone to selection bias, it was considered acceptable to continue with quasi-experimental studies. A total of 461 references were identified through both the database searches outlined above (n = 448) and from other sources (n = 13). Duplicates and irrelevant papers accounted for 432 references, leaving 29 references for full assessment. No further studies were identified by searching the references of the 29 papers obtained for full assessment. In total, 21 papers were excluded. The reasons for this are outlined in appendix 3. Eight studies meeting inclusion criteria were included in the qualitative synthesis (Bechtel et al., 2008, Campbell et al., 2012, Campbell et al., 2012a, Crandall, 2003, Derhammer et al., 2000, Kelly, 2004, Ledray and Simmelink, 1997, Sievers et al., 2003); six of these studies were also included in the meta-analysis (quantitative synthesis) (Bechtel et al., 2008, Campbell et al., 2012a, Crandall, 2003, Derhammer et al., 2000, Kelly, 2004, Ledray and Simmelink, 1997). The remaining two studies were included only in the qualitative synthesis; one because it presented the same data as another paper (Campbell et al., 2012), and one because it did not report data for any of the outcomes investigated in this review (Sievers et al., 2003). The flow of references is outlined in figure 1.

17 The Campbell Collaboration | www.campbellcollaboration.org

Figure 1. Reference Flow

4.1 DESCRIPTION OF STUDIES

(see characteristics of included studies table)

In total 2700 complainants were included in this meta-analysis. Of this group, 1223 complainants were cared for by a Sexual Assault Nurse Examiner (SANE), and 1477 were cared for by a non-SANE health professional. Non-SANE healthcare professionals included both specially trained and general physicians. Services in the UK tend to rely in the former, while those in the USA have tended to utilise general physicians where there is no SANE service. One study focussed on paediatric complainants of sexual assault and included a sample of 114 complainants (Bechtel et al., 2008). Ninety-eight per cent of the complainants in this study were female and the mean age was 14 years. One study evaluated the service offered to adult complainants by SANEs in the UK (Kelly, 2004) and included a sample of 1072 complainants, of whom 992 were female. All four remaining studies evaluated services offered to adult complainants in the USA. The sample sizes were 293 complainants (Campbell et al., 2012a); 957 female complainants, mean age 29 years (Crandall, 2003); 169 complainants (Derhammer et al., 2000); and 97 complainants (Ledray and Simmelink, 1997) respectively. Unless stated otherwise, mean age and gender of the complainants was not reported in the studies. Although multiple sites were utilised by some of the studies, all of the included sites provided both SANE and non-SANE services; some studies compared data collected both before and after

18 The Campbell Collaboration | www.campbellcollaboration.org

the implementation of a SANE service; others compared the provision of care to rape complainants when there was a SANE on shift to times when there was no SANE on shift.

4.2 RISK OF BIAS IN INCLUDED STUDIES

Of the eight studies included in this review, only one was assessed to be at low risk of selection bias (Bechtel et al., 2008). This study was considered to be at low risk of bias because it was a natural experiment with allocation to condition based on the department staff rota. That is, it was whether a SANE or non-SANE examiner was on duty at the time a complainant arrived for service. It is unlikely that complainant characteristics were related to staff rota and unlikely that staff were able to alter whether a SANE or non-SANE professional performed the examination. All complainants during the study period were included.

The remaining studies used a cohort or historical control group design. All complainants during a period of time when the hospital did not have a SANE examiner were compared to all complainants during a period of time when the hospital did have a SANE examiner.

The inclusion of all complainants during the study periods limits the potential for selection bias given that complainants do not self-select the timing of their victimisation and patient characteristics are unlikely to change systematically over time. However, it is entirely possible that there were historical changes that may have affected internal validity, such as changes to the hospital system or the criminal justice systems approach to handling sexual assault cases. No significant differences were noted between the time of day and treatment allocation across studies. None of the studies were able to blind participants and study personnel; however, one did refer to the blinding of outcome assessment (Bechtel et al., 2008). Only one study was judged to be free from attrition bias (Bechtel et al., 2008), and no study was without concern regarding either reporting or vested-interest bias. Figures 2 and 3 present a summary of the risks of bias across the included studies.

19 The Campbell Collaboration | www.campbellcollaboration.org

Figure 2. Risk of bias graph: investigators judgements about each risk of bias item presented as percentages across all included studies

Figure 3. Risk of bias summary: investigators judgements about each risk of bias item for each included study

20 The Campbell Collaboration | www.campbellcollaboration.org

4.3 EFFECTS OF INTERVENTION

The results from the meta-analysis are outlined in Tables 1 and 2 and the Summary of Findings Table. Ninety-five percent confidence intervals (95% CI) were stated for all outcomes and the summary measures used were risk ratio (RR) for the binary outcomes and mean difference (MD) for the continuous outcomes.

Complainants quality of life:

None of the included studies reported on this outcome.

Proportion of cases resulting in conviction:

This outcome was reported in two of the included studies (Campbell et al., 2012a, Kelly, 2004). No significant differences were found between the two intervention groups (RR 1.00, 95% CI 0.64 to 1.55; Figure 4).

Figure 4. SANE versus non-SANE conviction rate

Proportion of cases resulting in prosecution:

This outcome was reported in two of the included studies (Campbell et al., 2012a, Kelly, 2004). No significant difference was found between the SANE and non-SANE groups in terms of prosecution rate (RR 1.04, 95% CI 0.73 to 1.48; Figure 5)

Figure 5. SANE versus non-SANE prosecution rate

21 The Campbell Collaboration | www.campbellcollaboration.org

Complainant mortality within 30 days:

None of the included studies reported this outcome.

Time from complaint to examination:

None of the included studies reported this outcome.

Provision of STI, pregnancy and HIV prophylaxis:

STI prophylaxis: Two of the included studies contributed data to the STI prophylaxis outcome (Bechtel et al., 2008, Crandall, 2003). A small, yet statistically significant difference was reported, favouring complainants who were treated by a SANE, in terms of the provision of STI prophylaxis (RR 1.07, 95% CI 1.01 to 1.13; Figure 6).

Figure 6. SANE versus non-SANE provision of STI prophylaxis

Pregnancy prophylaxis: Two studies were included in the analysis of pregnancy prophylaxis (Bechtel et al., 2008, Crandall, 2003). As with STI prophylaxis, complainants cared for by a SANE were significantly more likely to be offered pregnancy prophylaxis than those cared for by a non-SANE health professional (RR 1.32, 95% CI 1.19 to 1.46; Figure 7).

Figure 7. SANE versus non-SANE provision of pregnancy prophylaxis

22 The Campbell Collaboration | www.campbellcollaboration.org

HIV prophylaxis: Only one of the included studies reported on the provision of HIV prophylaxis (Bechtel et al., 2008), and observed no significant difference between the two intervention groups in terms of the provision of HIV prophylaxis (RR 1.29; 95% CI 0.87 to 1.89).

Proportion of complainants who had a rape kit collected:

It was possible to extract data for this outcome from two of the included studies (Crandall, 2003, Derhammer et al., 2000). No significant differences were found for this outcome (RR 2.79, 95% CI 0.21 to 36.38; Figure 8a). However, when a fixed-effects model was applied, a significant difference was noted, favouring complainants who had been treated by a SANE (RR 3.94; 95% CI 3.21 to 4.84; figure 8b). It was also noted that, where a rape kit had been collected, the likelihood of the case being referred for prosecution was significantly increased (Campbell et al., 2012).

Figure 8a. SANE versus non-SANE proportion of complainants with rape kits collected (random-effects model)

Figure 8b. SANE versus non-SANE proportion of complainants with rape kits collected (fixed-effects model)

23 The Campbell Collaboration | www.campbellcollaboration.org

Proportion of complainants who had a forensic examination documented:

This outcome was reported in three of the included studies (Bechtel et al., 2008, Crandall, 2003, Derhammer et al., 2000), and no significant difference was found between the two intervention groups (RR 2.28, 95% CI 0.65 to 8.01; Figure 9a). However, when a fixed-effects model was employed, this difference was significant (RR 3.21, 95% CI 2.71 to 3.80; figure 9b). It was also noted in one study that documentation of the forensic examination was more complete or thorough in cases where treatment had been given by a SANE (Bechtel et al., 2008).

Figure 9a. SANE versus non-SANE proportion of complainants in whom a forensic examination was documented (random-effects model)

Figure 9b. SANE versus non-SANE proportion of complainants in whom a forensic examination was documented (fixed-effects model)

24 The Campbell Collaboration | www.campbellcollaboration.org

Number of rape kits admissible as evidence:

Only one of the included studies included data on this outcome (Ledray and Simmelink, 1997). This study found a statistically significant benefit to complainants cared for by SANEs in terms of the admissibility of their rape kit as evidence in court (RR 1.20, 95% CI 1.06 to 1.35). In addition to this, one of the studies included only for qualitative purposes did provide information regarding the accuracy of evidence collected by SANEs, compared with both doctors and non-SANE health professionals, with greater accuracy found in the case of forensic examinations conducted and reported by SANEs (Sievers et al., 2003). It was also reported that the rape kits were more complete and the chain of evidence less frequently broken when a SANE was involved in the case (Ledray and Simmelink, 1997, Sievers et al., 2003).

Cost per case:

Only one of the included studies documented this outcome (Kelly, 2004); however, as no standard deviation was reported, it is not possible to gauge the significance of the result. The results showed that utilising a non-SANE health professional in the provision of forensic examinations, in this case medical forensic doctors, costs were higher than when a SANE was utilised (non-SANEs 289 per case; SANEs 221 per case). The authors of this study did not indicate whether or not this difference was significant, nor did they provide information regarding the confidence intervals.

4.4 STATISTICAL VARIATIONS

Slight differences were noted between the fixed- and random-effects analyses; however, this only affected the significance of the results when considering the proportion of complainants for whom a rape kit was collected and the proportion for whom it was documented. No impact was noted when the effect measures were changed.

4.5 HETEROGENEITY

Significant statistical heterogeneity was noted in the following outcomes:

1. Proportion of cases resulting in prosecution (Chi = 3.16, df = 1 (P = 0.08); I = 68%).

2. Proportion of complainants who had a rape kit collected (Chi = 262.30, df = 1 (P < 0.00001); I = 100%).

3. Proportion of complainants who had a forensic examination documented (Chi = 170.22, df = 2 (P < 0.00001); I = 99%).

25 The Campbell Collaboration | www.campbellcollaboration.org

Due to the small number of studies included, and the high levels of bias reported across these studies, it was not possible to perform either subgroup or sensitivity analyses for any of these outcomes.

Proportion of cases resulting in prosecution:

With regard to the analysis of prosecution rates, the I would appear to suggest a significant element of heterogeneity in this outcome (I = 68%; P = 0.08), a fact which would seem to be reinforced by the fact that the two studies included (Campbell et al., 2012a, Kelly, 2004) are reporting conflicting results. However, the overlap of the confidence intervals of these two studies, suggests that the same effect estimate may be being measured in both studies. However, the wide confidence intervals reflect an element of variability within the samples, which may be accounted for by sampling error. Heterogeneity has also been observed in the proportion of cases resulting in conviction; however, this is not significant (Chi = 2.33, df = 1 (P = 0.13) I = 57%).

Proportion of complainants who had a rape kit collected:

It is worth noting that, individually, the papers included in the meta-analysis of the number of rape kits collected report a statistically significant result, favouring complainants treated by a SANE. The exclusion of either study from the analysis of the number of rape kits collected does reduce the magnitude of the effect, which suggests that while there is consistency across the results reported by the two papers, they may be recording and reporting this outcome in different ways. It is also worth noting that the confidence intervals of the two studies (Crandall, 2003, Derhammer et al., 2000) do not overlap, which also suggests that the populations sampled in the two studies may be different. This suggestion is reinforced by the relatively narrow confidence intervals, which suggest good degree of precision and low variability within the respective samples. It may also be possible that there are significant differences in treatment effects, comparators or the methodologies employed across the included studies.

Proportion of complainants who had a forensic examination documented:

With regard to the documentation of the forensic examination, all three papers reported a significant difference between the two treatment groups. However, the result from the Crandall study, more specifically the effect measure and confidence interval, does not overlap with those of the other two studies, which may imply that the population sampled in this study is not the same as the two remaining papers. However, the direction of the effect estimate is the same in all three studies, which suggests that, while the populations involved may be different, the effect remains consistent across the studies.

26 The Campbell Collaboration | www.campbellcollaboration.org

4.6 PUBLICATION BIAS

Due to the limited number of studies included within this meta-analysis, it was not possible to accurately detect the presence of publication bias.

4.7 SUMMARY OF FINDINGS

Sexual Assault Nurse Examiners compared to Physicians for the initial forensic examination of the victims of rape and sexual assault

Patient or population: patients with the initial forensic examination of the victims of rape and sexual assault Settings:

Intervention: Sexual Assault Nurse Examiners

Comparison: Physicians

Outcomes Illustrative comparative risks* (95% CI)

Relative effect (95% CI)

Comments

No of Participants (studies)

Quality of the evidence (GRADE)

Assumed risk Corresponding risk

Non-SANE health professionals

Sexual AssaultNurse Examiners

Conviction Rate Number of cases resulting in conviction

369 per 1000 369 per 1000 (236 to 573)

RR 1.00 (0.64 to 1.55)

435(2 studies)

very low1,2,3,4,5,6

Prosecution Rate Number of cases put forward for prosecution

484 per 1000 503 per 1000 (353 to 716)

RR 1.04 (0.73 to 1.48)

535(2 studies)

very low1,2,3,4,6,7

STI prophylaxis Proportion of patients offered STI prophylaxis

869 per 1000 930 per 1000 (878 to 982)

RR 1.07 (1.01 to 1.13)

888(2 studies)

very low1,2,3

Pregnancy prophylaxis Proportion of patients offered pregnancy prophylaxis

656 per 1000 866 per 1000 (780 to 957)

RR 1.32 (1.19 to 1.46)

922(2 studies)

very low1,2,3

HIV prophylaxis Proportion of patients offered HIV prophylaxis

583 per 1000 752 per 1000

(507 to 1000)

RR 1.29 (0.87 to 1.89)

60(1 study)

very low3,6,8,9

Rape kits collected Number of patients for whom a rape kit was collected

352 per 1000 982 per 1000

(74 to 1000)

RR 2.79 (0.21 to 36.38)

1126(2 studies)

very low1,2,3,10,11,12,13

Forensic examination documented

383 per 1000 872 per 1000

(249 to 1000)

RR 2.28 (0.65 to 8.01)

1240(3 studies)

very low1,2,3,13,14

27 The Campbell Collaboration | www.campbellcollaboration.org

Documentary evidence that the forensic examination took place

Kits admissible in a court of law Number of rape kits admissible as evidence in court

822 per 1000 986 per 1000 (871 to 1000)

RR 1.2 (1.06 to 1.35)

97(1 study)

very low1,3,4,6

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).

CI: Confidence interval; RR: Risk ratio;

GRADE Working Group grades of evidence

High quality: Further research is very unlikely to change our confidence in the estimate of effect.

Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.

Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.

Very low quality: We are very uncertain about the estimate.

1 No blinding

2 Incomplete outcome data

3 Selective reporting

4 Potential vested-interests

5 Total sample size is less than sample size calculated for a single trial using PS: Power and Sample Size calculation software (n = 1307)

6 Less than 300 events in total in both groups

7 Total sample size less than the sample size calculated for single trial using PS: Power and Sample Size calculation software (n = 617)

8 Only blinding of assessors

9 Total sample size less than that calculated for a single trial using PS: Power and Sample Size calculation software (n = 101)

10 No overlap of confidence intervals

11 Significant heterogeneity I = 100%; Chi = 262.3 (P < 0.00001)

12 Asymmetry on the funnel plot

13 Risk ratio greater than 2

14 Almost no overlap of confidence intervals and significant heterogeneity - I = 99%; Chi = 170.22 (P < 0.00001)

28 The Campbell Collaboration | www.campbellcollaboration.org

5 Discussion

This review has compared the overall reliability and efficacy of Sexual Assault Nurse Examiners (SANEs) versus non-SANE health professionals in the forensic examination of rape and sexual assault complainants. While data have been reported for adult services in both the UK and USA, paediatric cases have only been reported from the confines of a USA SANE programme (Bechtel et al., 2008). This may limit the generalisability of any findings pertaining to paediatric cases of rape and sexual assault to a UK population. Generalisability is also limited by the small number of relevant studies identified and the limited context and sample populations examined.

The most important outcome to consider is the quality of life of the complainants following both their rape and their treatment by either a physician or SANE. While it is unfortunate that none of the included studies reported on this, several service-user evaluations have been conducted, with the overall finding that a greater degree of control and empowerment was felt by complainants who were treated by a SANE (Campbell et al., 2008, Campbell et al., 2005). Also noted was a greater sense of empathy and feelings of support (Kelly, 2004). However, it is not clear how lasting these benefits may prove to be as no data are currently available regarding the longer-term effects on the complainants quality of life when treated by a SANE, compared with those treated by a non-SANE, so it may be that they are only short-lived.

No significant differences were found between SANEs and non-SANEs in terms of either conviction or prosecution rates. However, as the overall quality of the evidence has been graded as very low, it is not possible to state that the SANE intervention had no effect on these outcomes, but simply that there was a lack of evidence. It is also not possible to rule out the possibility of statistical heterogeneity, despite their overlapping confidence intervals, which implies that the same effect estimate is being measured in both studies. What is important to note is the width of the confidence intervals. In both studies, these are quite wide, suggesting a large element of variability within the sample populations, which may reflect a degree of sampling error. This, consequently, reduces the power or precision of the point estimate and indicates heterogeneity within the sample. Also, as the direction of the effect is different in the two studies, this suggests that the outcomes (conviction and prosecution) are not consistent across the studies. However, while this is not

29 The Campbell Collaboration | www.campbellcollaboration.org

significant in terms of the conviction rates (P = 0.13), this does appear to be significant when considering the prosecution rates (P = 0.08). As both papers compare services offered to adult complainants, it would be reasonable to assume that differences in results could be accounted for by the differences in the services themselves, and may be tied into the fact that one paper has evaluated the service offered to rape complainants in the UK (Kelly, 2004), while the other has focused on the USA (Campbell et al., 2012b). It may also be that these conflicting results reflect differences in the UK and USA penal systems, or significant differences in the UK and USA control groups. In the UK, non-SANE healthcare professionals are typically highly trained forensic medical examiners. However, in the USA, non-SANE healthcare professionals are typically general emergency room physicians. Also, whereas the UK SANEs have no involvement in the prosecution and conduct of any criminal proceedings, their counterparts in the USA have a much greater and continued involvement with these cases. Indeed, within the USA courtrooms, the SANE is a well-accepted and respected source of expert testimony. This alone may lend greater credibility to the complainants of rape and sexual assault and, consequently, enhance the probability of any case moving forward to prosecution and conviction.

In terms of the clinical care of the complainants, SANEs were significantly more likely than physicians to offer both STI and pregnancy prophylaxis. No significant differences were found in terms of HIV prophylaxis. It is not clear whether these differences were due to the specialised training offered to the forensic nursing staff or simply that the nursing specialists were able to spend more time with the complainants. Also, the fact that SANEs are highly specialised and specifically trained to deal with complainants of rape and sexual assault, is likely to increase the possibility of their offering the most thorough and appropriate care to these individuals. Placing this in the context of the UK findings on the rates of conviction and prosecution, there seems to be a decision to be made in terms of care or criminal justice. While no significant differences were noted in terms of conviction or prosecution, in terms of clinical care, the complainants would seem to be offered a more comprehensive service if a SANE is involved in their case.

Regarding professional practice, the findings are dependent on the method of meta-analysis employed. Using the random-effects model, no significant differences were found in terms of the collection and documentation of the forensic examination when treatment is given by a SANE or a non-SANE health professional. However, using the fixed-effects model, the complainants were more likely both to have a rape kit collected, and to have this documented when a SANE was involved in their case. This is important, not only in terms of continuity of care and best practice, but also because cases are more likely to progress to conviction when a forensic examination has been documented (Bechtel et al., 2008). The accuracy of the forensic evidence collection process and, more specifically, the proportion of rape kits that can be used as evidence in court is also important. While only one paper reported this outcome

30 The Campbell Collaboration | www.campbellcollaboration.org

specifically (Ledray and Simmelink, 1997), the results suggest that when a rape kit has been collected by a SANE, it is significantly more likely to be admissible in court than those collected by a non-SANE health professional. This finding is indirectly supported by another study which has found that SANEs are more accurate, thorough, and less likely to breach the chain of evidence (Sievers et al., 2003).

With this in mind, it may well be that this is another factor to consider when addressing the current prosecution and conviction rates. It would appear that, not only does best practice need to be established, particularly in the UK, it should also be championed and enforced, to ensure the best possible legal and clinical outcomes for the complainants of rape and sexual assault. Also, while establishing the best approach to complainants of rape or sexual assault; it is worth noting that SANEs are, on average, 68 per case cheaper than their physician counterparts.

It may be worth considering at this point, the potential negative impacts of the increased reliance on SANEs. While this is not currently a major issue within the UK, the SANE programmes in the USA and Canada have now been operational for well over 30 years, and their impact has been noted. By far the most important factor to consider is the consequent de-skilling of doctors within the field of forensic examination. Indeed, many emergency department physicians are reluctant to conduct these examinations, feeling that they lack the specific training and experience to carry them out correctly (Campbell et al., 2012, Ledray and Simmelink, 1997). In cases where an emergency department physician conducts the forensic examination, this de-skilling may lead to the rape kit being collected incorrectly, a fact which has, to some extent been illustrated by the meta-analysis reported here (Ledray and Simmelink, 1997, Sievers et al., 2003). Also, a poorly conducted forensic medical examination may result in the complainant feeling revictimised as a result of their negative experience. This, in turn, may increase the likelihood of the complainant choosing to withdraw their rape complaint (Campbell et al., 2012), which will likely impact negatively on both the proportion of cases being prosecuted and those resulting in conviction.

The withdrawal of rape complaints is a key factor to consider when investigating conviction rates because sexual assault cases have the highest attrition and lowest conviction rates of all serious offences (Lea et al., 2003). This may be reflected by the fact that, although the reporting of rape and other forms of sexual violence has increased, the actual conviction statistics have seen little change (Kelly et al., 2005). Attrition is the process whereby cases drop out of the criminal justice system at one of a number of points of exit from that system (Lea et al., 2003). These stages include the decision to report the offence, the investigation stage, discontinuance by the prosecutorial team, and acquittal or late withdrawal at the trial stage (Kelly et al., 2005, Lea et al., 2003). Up to two thirds of cases are dropped or drop out in the investigative stage. This includes withdrawal by the complainants, withdrawal due to evidential issues, and issues surrounding complainant credibility. As a result, it is

31 The Campbell Collaboration | www.campbellcollaboration.org

estimated that only around 14% of cases actually reach trial stage (Kelly et al., 2005). It may be that the implementation of a SANE-led service, which has been shown here to offer a more comprehensive level of care to the complainants of rape and sexual assault, may lead to an increase in the proportion of cases being reported, prosecuted and those resulting in conviction.

Returning to the initial question surrounding the overall efficacy and reliability of SANEs compared to non-SANE health professionals, it is important to consider the potential and desired consequences and ramifications of the substitution of forensic doctors with SANEs. In terms of conviction rates alone one could, perhaps, assume that a better service or better complainant experience could, in addition to improving the quality of the evidence gathered, lead to a greater willingness on the part of the complainant to both report and proceed with their case through the criminal justice system. Additionally, one would hope that by maintaining and improving links between the various agencies involved in such cases, we could instil a greater faith, on the part of the complainant, in the police force, Criminal Prosecution Service (CPS), and judicial system. This could, in turn, increase the number of complainants reporting and proceeding with their cases to trial.

With regard to the actual conduct and methodology of this review, it is worth noting that, despite a thorough and extensive search of all relevant databases and grey literature, no randomised controlled trials were identified, and those papers that were considered relevant only numbered eight. Also of note is the fact that only one of the six studies included in the meta-analysis reported data from the UK. While this sample population represented 56% of the complainants treated by a non-SANE health professional, only 20% of the SANE sample was accounted for by this study. This large imbalance in the sample sizes and the questions surrounding practice may undermine the generalisability of the findings of this review. It also calls into question whether the UK service is actually comparable to that in the USA. The significant heterogeneity reported in terms of the proportion of cases resulting in prosecution would seem to suggest not and it is not clear if the source of the heterogeneity is the SANE service itself, or fundamental differences between the UK and USA judicial systems. It is also important to consider the internal validity of the included studies, particularly those which employed a before-after study design. It may well be that, in addition to the change from doctor to SANE, there were also changes to hospital policy, facility provision and the judicial system.

In considering the generalisability of the results of this review to a UK population, it is worth investigating the potential differences between UK and USA services. While both SANEs and FNEs were introduced as a potential solution to the problem of recruiting and retaining female staff, the SANE programmes in the USA are well established and have been in operation for well over 30 years (Ledray and Simmelink, 1997); however, FNEs in the UK are relatively scarce. Indeed, the role itself has only been in development in the UK since around 2000 (Rutty, 2006). As

32 The Campbell Collaboration | www.campbellcollaboration.org

such, their services are not in routine use throughout the entire UK. This poses two important questions:

1. Are the USA SANEs better trained than the UK FNEs? 2. Do the two programmes operate in the same manner?

In the USA, nurses are required to complete in excess of 40 hours of classroom-based training, followed by 40 to 90 hours of clinical training prior to qualifying as a SANE (Campbell et al., 2012a, Sievers et al., 2003). However, in the UK there is currently no standardised training programme available to those wishing to specialise in forensic nursing. In addition to the potential impact of this factor on the skills and abilities of the FNEs, this may also lead to the FNEs having their credibility and capabilities questioned by the police, the complainants of rape and the CPS. This may, consequently, impact negatively on the prosecution and conviction rates in cases involving a FNE. This being said, this review has found no significant difference in rates of conviction and prosecution between complainants treated by FNEs and those treated by forensic doctors.

In the circumstances, the best available evidence from both published and unpublished sources was included. Also, the papers were selected by two independent reviewers, to ensure that all relevant studies were identified and included. The collaboration of two independent reviewers also ensured that, as far as possible, the studies provided relevant data, and that it was extracted correctly. Unfortunately it is not clear if the results reported here can be generalised to the UK because some outcome measures suggest differences between the UK and USA, whereas others suggest no significant differences. The fact that, comparatively speaking, the UK service is still very much in its infancy, whereas that in the USA is reasonably well established, may provide one reason for the reported differences and lack of available evidence and research. Another may be the differences in the UK and USA approaches to crime and justice. With this in mind, one could argue that the combination of data from these two very different services is inappropriate. It is entirely plausible to suggest that the heterogeneity between the two services is the main cause of the equivocal results reported here. However, in the absence of further research, the decision was made to proceed with all available data.

While the lack of randomised studies and lack of work within a UK population may limit the generalisability of the conclusions of this review, a number of key findings need to be borne in mind. Most importantly that the implementation of a nurse-led service does not seem to have a negative impact on prosecution and conviction rates and, in a majority of cases, leads to better provision of care, in terms of the STI and pregnancy prophylaxis offered to the complainants. However, perhaps the most important element of all has been missed here, that of the complainant him-/herself. Indeed, as Baroness Stern reports (Stern, 2010, pp101):

33 The Campbell Collaboration | www.campbellcollaboration.org

By having an obsession with the attrition and conviction rates we are putting the criminal justice system at the centre, not the victim.

This is a key point to consider; particularly in light of the finding that the vast majority of women feel that conviction is far less important than their overall treatment throughout the process of complaint, investigation, and prosecution (Stern, 2010).

34 The Campbell Collaboration | www.campbellcollaboration.org

6 Conclusions

6.1 IMPLICATIONS FOR PRACTICE

While there is a lack of evidence of effect in terms of conviction and prosecution rates, this review demonstrates the superiority of SANEs/FNEs in terms of clinical care, more specifically the provision of STI and pregnancy prophylaxis. This, taken in tandem with the fact that a FNE is around 70 per case cheaper than a forensic physician, would seem to lend strong support to further investigation of the ongoing role of forensic nurses within the field of sexual violence.

6.2 IMPLICATIONS FOR RESEARCH

Research to investigate the quality of life of the complainants following their rape and forensic examination, both in the short and longer term is needed. This work should encompass not only physical aspects of the complainants quality of life, but also any on-going emotional and psychological issues which may arise as a result, not only of the rape itself, but also as a consequence of the secondary victimisation experienced by many as a consequence of the forensic examination, police investigation and legal proceedings.

Studies evaluating the overall quality and efficiency of nurse- and doctor-led services in terms of time from complaint to examination, collection and documentation of rape kits and the admissibility of the evidence gathered from such kits should be conducted on a much larger scale than is currently the case. It is hoped that the results from studies of this type, when combined with those pertaining to the complainants quality of life will provide more robust and conclusive evidence to either support of refute the implementation of a FNE-led service for the complainants of sexual assault.

Research to establish the barriers to the implementation of a nurse-led service for the forensic examination of the complainants of rape and sexual assault, along with a full economic analysis of the nurse- and physician-led services is also necessary, particularly with reference to a UK population.

35 The Campbell Collaboration | www.campbellcollaboration.org

7 References

DEMETS, D. L. 1987. Methods for combining randomized clinical trials: Strengths and limitations. Statistics in Medicine, 6, 341-348.

DEPARTMENT OF HEALTH, HOME OFFICE & ASSOCIATION OF CHIEF POLICE

OFFICERS. 2009. Revised National Service Guide: a resource for developing Sexual Assault Referral Centres.

DERSIMONIAN, R. & LAIRD, N. 1986. Meta-analysis in clinical trials. Controlled

Clinical Trials, 7, 177-188.

DU MONT, J. D., MILLER, K.-L. & MYHR, T. L. 2003. The Role of "Real Rape" and

"Real Victim" Stereotypes in the Police Reporting Practices of Sexually Assaulted Women. Violence Against Women, 9, 466-486.

DUBOURG, R. & ET AL. 2005. The economic and social costs of crime against individuals and households 2003/04 (Home Office online report 30/05).

NHS EMPLOYERS. 2012. Pay Circular (M & D) 1/2012 [Online]. Available: http://www.nhsemployers.org/Aboutus/Publications/Documents/Pay-Circular-MD-1-2012.pdf [Accessed 13/08/2012].

H M GOVERNMENT. Cross-government Action Plan on Sexual Violence and Abuse

[Online]. Available: http://webarchive.nationalarchives.gov.uk/+/http://www.homeoffice.gov.u k/documents/Sexual-violence-action-plan2835.pdf?view=Binary.

HIGGINS, J. & GREEN, S. 2011. Cochrane Handbook for Systematic Reviews of interventions Version 5.1.0 [updated March 2011]. The Cochrane Collaboration.

HIGGINS, J. P. T. & THOMPSON, S. G. 2002. Quantifying heterogeneity in a meta-analysis. Statistics in Medicine, 21, 1539-1558.

36 The Campbell Collaboration | www.campbellcollaboration.org

HOME OFFICE, DEPARTMENT OF HEALTH, CARE SERVICES IMPROVEMENT

PARTNERSHIP & MENTAL HEALTH IN ENGLAND. 2005. National service guidelines for developing sexual assault referral centres (SARCs).

HOME OFFICE. 2003. Sexual Offences Act [Online]. Available: http://www.legislation.gov.uk/ukpga/2003/42/contents.

INTERNATIONAL ASSOCIATION OF FORENSIC NURSES. 2006. The role of

SANEs [Online]. Available: http://www.iafn.org/displaycommon.cfm?an=1&subarticlenbr=546.

INOTT, D. 22 March 2012 2012. RE: Personal communication. Type to TOON, C. D.

KELLY, L., LOVETT, J. & REGAN, L. 2005. A Gap or A Chasm?: Attrition in

Reported Rape Cases. A Gap or A Chasm?: Attrition in Reported Rape Cases, 136, 136 pp.

LEA, S. J., LANVERS, U. & SHAW, S. 2003. Attrition in rape cases: Developing a profile and identifying relevant factors. British Journal of Criminology, 43, 583-599.

NATIONAL SOCIETY FOR THE PREVENTION OF CRUELTY TO CHILDREN.

2012. Statistics on Child Sexual Abuse [Online]. Available: http://www.nspcc.org.uk/Inform/resourcesforprofessionals/sexualabuse/statistics_wda87833.html

Web End =http://www.nspcc.org.uk/Inform/resourcesforprofessionals/sexualabuse/st http://www.nspcc.org.uk/Inform/resourcesforprofessionals/sexualabuse/statistics_wda87833.html

Web End =atistics_wda87833.html . [Accessed 27/07/2012.

ROYAL COLLEGE OF NURSING. 2012. Agenda for Change pay bands effective from 1 April 2011 [Online]. Available: http://www.rcn.org.uk/support/pay_and_conditions/pay_rates_20112012 [Accessed 13/08/2012].

RUTTY, J. E. 2006. Does England Need a New Genesis of Forensic Nursing?

Forensic Science, Medicine, and Pathology, 2, 149-155.

STERN, B. V. 2010. A report by Baroness Vivien Stern CBE of an independent review into how rape complainants are handled by publich authorities in England and Wales. London: Home Office.

UNITED KINGDOM ASSOCIATION OF FORENSIC NURSES. 2012. SARC Map

[Online]. Available: http://www.ukafn.org/map/ [Accessed 27/07/2012.

UNITED KINGDON ASSOCIATION OF FORENSIC NURSES. 2012a. About UKAFN

[Online]. Available: http://www.ukafn.org/ [Accessed 24/09/2012.

37 The Campbell Collaboration | www.campbellcollaboration.org

8 References for eligible studies

BECHTEL, K., RYAN, E. & GALLAGHER, D. 2008. Impact of sexual assault nurse examiners on the evaluation of sexual assault in a pediatric emergency department. Pediatric Emergency Care, 24, 442-447.

CAMPBELL, R., BYBEE, D., KELLEY, K. D., DWORKIN, E. R. & PATTERSON, D.

2012. The Impact of Sexual Assault Nurse Examiner (SANE) Program services on law enforcement investigational practices: A mediational analysis. Criminal Justice and Behavior, 39, 169-184.

CAMPBELL, R., PATTERSON, D. & BYBEE, D. 2012a. Prosecution of adult sexual assault cases: a longitudinal analysis of the impact of a sexual assault nurse examiner program. Violence Against Women, 18, 223-44.

CRANDALL, C. S. H., D 2003. Impact Evaluation of a Sexual Assault Nurse

Examiner (SANE) Program. United States: Department of Justice.

DERHAMMER, F., LUCENTE, V., REED 3RD, J. F. & YOUNG, M. J. 2000. Using a

SANE interdisciplinary approach to care of sexual assault victims. Jt Comm J Qual Improv, 26, 488-496.

KELLY, L. L., J. REGAN, L 2004. Forensic nursing: an option for improving responses to reported rape and sexual assault.

LEDRAY, L. E. & SIMMELINK, K. 1997. Efficacy of SANE evidence collection: A

Minnesota study. J Emerg Nurs, 23, 75-7.

SIEVERS, V., MURPHY, S. & MILLER, J. J. 2003. Sexual assault evidence collection more accurate when completed by sexual assault nurse examiners: Colorado's experience. J Emerg Nurs, 29, 511-4.

38 The Campbell Collaboration | www.campbellcollaboration.org

9 References for excluded studies

2002. SANE (sexual assault nurse examiner) programs have benefits you can't ignore. ED Manag, 14, 64-5.

AHRENS, C. E., CAMPBELL, R., WASCO, S. M., APONTE, G., GRUBSTEIN, L. &

DAVIDSON, W. S. 2000. Sexual assault nurse examiner (SANE) programs - Alternative systems for service delivery for sexual assault victims. J Interpers Violence, 15, 921-943.

AIKEN, M. M. & SPECK, P. M. 1995. Sexual assault and multiple trauma: a sexual assault nurse examiner (SANE) challenge. JEN: Journal of Emergency Nursing, 21, 466-469.

BERNARD, D., HARTZ, G., FARNER, K., KING, W. & MONROE, K. 2005.

Reliability and review of pediatric sexual assault nurse examiner findings. Journal of Investigative Medicine, 53, S323-S323.

BRAYDEN, R. M., ALTEMEIER, I. W. A., YEAGER, T. & MURAM, D. 1991.

Interpretations of colposcopic photographs: Evidence for competence in assessing sexual abuse? Child Abuse and Neglect, 15, 69-76.

BURGESS, A. W., LEWIS-O'CONNOR, A., NUGENT-BORAKOVE, M. & FANFLIK, P. 2006. SANE/SART Services for Sexual Assault Victims: Policy Implications. Victims & Offenders, 1, 205-212.

CAMPBELL, R., GREESON, M. R., BYBEE, D. & FEHLER-CABRAL, G. 2011.

Adolescent Sexual Assault Victims and the Legal System: Building Community Relationships to Improve Prosecution Rates. Am J Community Psychol.

CAMPBELL, R., PATTERSON, D., ADAMS, A. E., DIEGEL, R. & COATS, S. 2008. A participatory evaluation project to measure SANE nursing practice and adult sexual assault patients' psychological well-being. Journal of Forensic Nursing, 4, 19-28.

39 The Campbell Collaboration | www.campbellcollaboration.org

CAMPBELL, R., PATTERSON, D. & BYBEE, D. 2011a. Using Mixed Methods to

Evaluate a Community Intervention for Sexual Assault Survivors: A Methodological Tale. Violence Against Women.

CAMPBELL, R., PATTERSON, D., BYBEE, D. & DWORKIN, E. R. 2009. Predicting sexual assault prosecution outcomes: The role of medical forensic evidence collected by sexual assault nurse examiners. Criminal Justice and Behavior, 36, 712-727.

CAMPBELL, R., PATTERSON, D. & FEHLER-CABRAL, G. 2010. Using Ecological Theory to Evaluate the Effectiveness of an Indigenous Community Intervention: A Study of Sexual Assault Nurse Examiner (SANE) Programs. American Journal of Community Psychology, 46, 263-276.

CAMPBELL, R., PATTERSON, D. & LICHTY, L. F. 2005. The effectiveness of sexual assault nurse examiner (SANE) programs: a review of psychological, medical, legal, and community outcomes. Trauma Violence Abuse, 6, 313-330.

CHRISTOFIDES, J., JEWKES, K., WEBSTER, N., PENN-KEKANA, L., ABRAHAMS, N. & MARTIN, J. 2007. "Other patients are really in need of medical attention" -- the quality of health services for rape survivors in South Africa. Bulletin of the World Health Organization, 85, 495-503.

CHU, J. J. 2004. Sexual assault nurse examiners - The evidence collection kits they prepare are more accurate and complete. American Journal of Nursing, 104, 72AA-72AA.

FULGINITI, T. P., SEIBERT, E., FIRTH, V., QUATTRONE, J., HARNER, L., WEBB, J. & VIDA, P. 1996. A SANE experience in a community hospital: Doylestown's first six months... Sexual Assault Nurse Examiner (SANE). JEN: Journal of Emergency Nursing, 22, 422-426.

LEDRAY, L. 2005. Data on SANE programs crucial to refining the specialty, improving care. Journal of Forensic Nursing, 1, 187-188.

LEWIS-O'CONNOR, A. 2009. The evolution of SANE/SART--Are there differences?

Journal of Forensic Nursing, 5, 220-227.

PATTERSON, D. & CAMPBELL, R. 2012. The Problem of Untested Sexual Assault Kits: Why Are Some Kits Never Submitted to a Crime Laboratory? J Interpers Violence.

40 The Campbell Collaboration | www.campbellcollaboration.org

PENNINGTON, E. C., ZWEMER JR, F. L. & KREBS, D. A. 2010. Unique Sexual Assault Examiner Program Utilizing Mid-Level Providers. Journal of Emergency Medicine, 38, 95-98.

PLICHTA, S. B., CLEMENTS, P. T. & HOUSEMAN, C. 2007. Why SANEs matter: models of care for sexual violence victims in the emergency department. Journal of Forensic Nursing, 3, 15-23.

STERMAC, L., DUNLAP, H. & BAINBRIDGE, D. 2005. Sexual assault services delivered by SANEs. J Forensic Nurs, 1, 124-8.

THOMAS, M. & ZACHRITZ, H. 1993. Tulsa Sexual Assault Nurse Examiners (SANE) Program. The Journal of the Oklahoma State Medical Association, 86, 284-6.

41 The Campbell Collaboration | www.campbellcollaboration.org

10 Tables

Data and analysis:

Table 1a. Summary of meta-analysis (random effects model)

Outcome Studies Participants Statistical Method Effect Estimate

Conviction rate 2 435 RR (M-H, random, 95% CI) 1.00 (0.64, 1.53)

Prosecution rate 2 535 RR (M-H, random, 95% CI) 1.04 (0.73, 1.48)

STI prophylaxis 2 888 RR (M-H, random, 95% CI) 1.07 (1.01, 1.13)

Pregnancy prophylaxis

2 922 RR (M-H, random, 95% CI) 1.32 (1.19, 1.46)

HIV prophylaxis 1 60 RR (M-H, random, 95% CI) 1.29 (0.87, 1.89)

Kit collected 2 1126 RR (M-H, random, 95% CI) 2.79 (0.21, 3.38)

Kit documented 3 1240 RR (M-H, random, 95% CI) 2.28 (0.65, 8.01)

Kit admissible 1 97 RR (M-H, random, 95% CI) 1.20 (1.06, 1.35)

Cost per case 1 496 MD (IV, random, 95% CI) Not estimable

42 The Campbell Collaboration | www.campbellcollaboration.org

Table 1b. Summary of meta-analysis (fixed effects model)

Outcome Studies Participants Statistical Method Effect Estimate

Conviction rate 2 435 RR (M-H, fixed, 95% CI) 1.04 (0.78, 1.39)

Prosecution rate 2 535 RR (M-H, fixed, 95% CI) 1.05 (0.86, 1.29)

STI prophylaxis 2 888 RR (M-H, fixed, 95% CI) 1.07 (1.01, 1.13)

Pregnancy prophylaxis

2 922 RR (M-H, fixed, 95% CI) 1.32 (1.19, 1.46)

HIV prophylaxis 1 60 RR (M-H, fixed, 95% CI) 1.29 (0.87, 1.89)

Kit collected 2 1126 RR (M-H, fixed, 95% CI) 3.94 (3.21, 4.84)

Kit documented 3 1240 RR (M-H, fixed, 95% CI) 3.21 (2.71, 3.80)

Kit admissible 1 97 RR (M-H, fixed, 95% CI) 1.20 (1.06, 1.35)

Cost per case 1 496 MD (IV, fixed, 95% CI) Not estimable

43 The Campbell Collaboration | www.campbellcollaboration.org

44 The Campbell Collaboration | www.campbellcollaboration.org

Characteristics of included studies:

Bechtel 2008

Methods Cohort study

Participants Menarchal complainants and male adolescents with a history of sexual assault involving contact with the alleged perpetrators genitalia or with biological fluids from the alleged perpetrator, presented for medical evaluation, and requiring a forensic examination (n = 114).

Interventions Paediatric SANE (n = 60). Paediatric emergency department physicians (n = 54). Outcomes STI, pregnancy and HIV prophylaxis.

Documentation of forensic examination. Time from assault to presentation. Notes

Risk of bias

Bias Authors judgement

High risk Blinding is unlikely to have occurred.

Low risk The outcome assessors were blinded to treatment allocation.

Low risk A consecutive series of complainants was included.

High risk Important outcomes were not reported.

Vested-interest bias Unclear risk Some of the authors are likely to be SANEs. It would be in the investigators best-interests to demonstrate the success and efficacy of their newly-established programme.

Campbell 2012

Methods Cohort study. Data for the control group were collected prior to the implementation of the SANE programme; the intervention data after.

Participants Sexual assault complainants meeting the following criteria: 1. The reported crime was classified as a criminal sexual conduct offense.

2. The complainant was at least 18 years old.

45 The Campbell Collaboration | www.campbellcollaboration.org

Support for judgement

Selection bias Low risk This was a natural experiment, with group allocation dependent on the duty rota for SANEs and doctors. No significant differences were noted between the time of day and treatment allocation.

Blinding of participants and personnel (performance bias)

Blinding of outcome assessment (detection bias)

Incomplete outcome data (attrition bias)

Selective reporting (reporting bias)

3. The crime occurred between January 1994 and December 2005.

Interventions SANE. Emergency department personnel prior to the implementation of the SANE programme.

Outcomes Prosecution. Documentation of the forensic examination. Notes Included only in the qualitative synthesis.

Risk of bias

Bias Authors judgement

High risk Blinding not possible.

High risk Blinding not possible.

High risk Data were not available for a significant number of pre-SANE cases.

High risk Important outcomes were not reported.

Vested-interest bias High risk The main author performs collaborative research with SANE programmes.

Support for judgement

Selection bias Unclear risk This was a natural experiment. Data were compared from participants before and after the implementation of the SANE programme. While there are not likely to have been changes to participant characteristics over time, it is not possible to rule out the potential for changes to both the hospital and/or prosecutorial systems.

Blinding of participants and personnel (performance bias)

Blinding of outcome assessment (detection bias)

Incomplete outcome data (attrition bias)

Selective reporting (reporting bias)

46 The Campbell Collaboration | www.campbellcollaboration.org

Campbell 2012a

Methods Cohort study.

Participants Inclusion criteria: 1. Adult sexual assault cases. 2. Complainant assaulted within the focal county. 3. Case investigated by one of the five largest police departments in the county.

4. A complete forensic medical examination conducted by either a SANE or county hospital personnel.

5. Examination results analysed by the state crime lab for DNA evidence.

Exclusion criteria: 1. Complainant charged with false reporting. 2. Complainant retracted their statement. 3. Offender not identified. 4. Outcomes not available.

Interventions SANE available (24/7) in a facility outside the hospital (n = 137). County hospital personnel, prior to the implementation of the SANE programme (n = 156).

Outcomes Prosecution. Conviction. Notes

Risk of bias

Bias Authors judgement

High risk Blinding was not possible.

High risk Blinding was not possible.

High risk A number of eligible complainants were excluded.

High risk Important outcomes were not reported.

Vested-interest bias High risk Themain author performs collaborative research with SANE programmes.

47 The Campbell Collaboration | www.campbellcollaboration.org

Support for judgement

Selection bias Unclear risk This was a natural experiment. Data were compared from participants before and after the implementation of the SANE programme. While there are not likely to have been changes to participant characteristics over time, it is not possible to rule out the potential for changes to both the hospital and/or prosecutorial systems.

Blinding of participants and personnel (performance bias)

Blinding of outcome assessment (detection bias)

Incomplete outcome data (attrition bias)

Selective reporting (reporting bias)

Crandall 2003

Methods Cohort study. Data for the control group were collected prior to the implementation of the SANE programme; the intervention data after.

Participants Data for female complainants of sexual assault who were aged 18 years on January 1st 2000, who presented for forensic examination from 1994 to 1999.

Interventions SANE (n = 715). Emergency department staff prior to the implementation of the SANE programme (n = 242).

Outcomes STI and pregnancy prophylaxis. Number of rape kits collected. Documentation of the forensic examination. Notes

Risk of bias

Bias Authors judgement

High risk Blinding not possible.

High risk Blinding not possible.

High risk Several complainants were excluded from various outcomes.

High risk Important outcomes were not reported.

Vested-interest bias Unclear risk Funded by a National Institute of Justice grant. A poor outcome may reflect badly on the decision to fund SANE programmes.

Derhammer 2000

Methods Cohort study. Data for the control group were collected prior to the implementation of the SANE programme; the intervention data after.

Participants Rape complainants before and after the implementation of the

SANE programme.

Interventions SANE (n = 39). Medical examiners prior to the implementation of the

48 The Campbell Collaboration | www.campbellcollaboration.org

Support for judgement

Selection bias Unclear risk This was a natural experiment. Data were compared from participants before and after the implementation of the SANE programme. While there are not likely to have been changes to participant characteristics over time, it is not possible to rule out the potential for changes to both the hospital and/or prosecutorial systems.

Blinding of participants and personnel (performance bias)

Blinding of outcome assessment (detection bias)

Incomplete outcome data (attrition bias)

Selective reporting (reporting bias)

SANE programme (n = 130).

Outcomes Number of rape kits collected. Documentation of the forensic examination. Notes

Risk of bias

Bias Authors judgement

High risk Blinding not possible.

High risk Blinding not possible.

Unclear risk Not reported.

High risk Important outcomes were not reported.

Vested-interest bias Unclear risk Not reported.

Kelly 2004

Methods Cohort study.

Participants Sexual assault complainants presenting between 1st October 2000 and 31st December 2002.

Interventions Forensic nurse examiners (UK equivalent of SANE; n =

249). Forensic medical doctor (n = 822). Group assignment was based on which staff were available at the time the complainant was presented for examination

Outcomes Conviction. Prosecution. Notes

Risk of bias

Bias Authors judgement

49 The Campbell Collaboration | www.campbellcollaboration.org

Support for judgement

Selection bias Unclear risk This was a natural experiment. Data were compared from participants before and after the implementation of the SANE programme. While there are not likely to have been changes to participant characteristics over time, it is not possible to rule out the potential for changes to both the hospital and/or prosecutorial systems.

Blinding of participants and personnel (performance bias)

Blinding of outcome assessment (detection bias)

Incomplete outcome data (attrition bias)

Selective reporting (reporting bias)

Support for judgement

Selection bias High risk One nurse versus 13 doctors. Also the nurse examined more complainants during the daytime than the doctors. This could have affected the

overall effect estimate.

Blinding of participants and personnel (performance bias)

High risk Blinding not possible.

Blinding of outcome assessment (detection bias)

High risk Blinding not possible.

Incomplete outcome data (attrition bias)

High risk Not all records contained a complete set of data and those not undergoing forensic examination were excluded,

High risk Important outcomes were not reported.

Vested-interest bias Unclear risk The study was funded by the Crime Reduction Programme, Violence Against Women initiative. It is not clear what impact a negative outcome may have had.

Ledray 1997

Methods Cohort study.

Participants Rape complainants who had evidence kits sent to the forensic lab. Kits were analysed between February and October 1996.

Interventions SANE (n = 24). Non-SANE and physicians (n = 73) Outcomes Proportion of rape kits admissible as evidence in court.

Notes

Risk of bias

Bias Authors judgement

Support for judgement

Selection bias Unclear risk The selection process was not reported.

Blinding of participants and personnel (performance bias)

High risk No blinding.

Selective reporting (reporting bias)

Blinding of outcome assessment (detection bias)

High risk No blinding.

Incomplete outcome data (attrition bias)

Unclear risk The authors reported requesting an audit of rape kits, yet only 97 have been presented in this report.

Selective reporting (reporting bias)

High risk Important outcomes were not reported.

Vested-interest bias High risk One of the authors is a SANE.

50 The Campbell Collaboration | www.campbellcollaboration.org

Sievers 2003

Methods Cohort study.

Participants Rape complainants who had a forensic examination and kit sent to one of three participating forensic labs. Kits submitted between October 1999 and April 2002.

Interventions SANE (n = 276). Non-SANE and physicians (n = 236). Outcomes Accuracy of evidence collection.

Notes Included only in the qualitative synthesis.

Risk of bias

Bias Authors judgement

Support for judgement

Selection bias Unclear risk The selection process was not reported.

Blinding of participants and personnel (performance bias)

High risk Blinding unlikely.

Blinding of outcome assessment (detection bias)

High risk Blinding unlikely.

Incomplete outcome data (attrition bias)

Unclear risk A full rape kit was not collected for all complainants.

Selective reporting (reporting bias)

High risk Important outcomes were not reported.

Vested-interest bias High risk The main author is a SANE.

51 The Campbell Collaboration | www.campbellcollaboration.org

Characteristics of excluded studies:

Study ID Reason for exclusion

Ahrens 2000 Head of programme interviewed

Aiken 1995 Case report

Bernard 2005 No control group

Brayden 1991 Nurses/doctors surveyed

Burgess 2006 No separate data for SANE

Campbell 2005 Review

Campbell 2009 No control group

Campbell 2010 Police/prosecutors surveyed

Campbell 2011 No control group

Campbell 2011a Non-comparative study

Christofides 2007 Doctors/nurses surveyed

Chu 2004 Comment on Sievers 2003

Fulginiti 1996 Non-comparative study

Ledray 2005 Non-comparative study

Lewis-OConnor 2009 No separate SANE data

New study Personal opinion

Patterson 2012 No control group

Pennington 2010 Paramedics utilised

Plichta 2007 Emergency department survey

Stermac 2005 No control group

Thomas 1993 Non-comparative study

52 The Campbell Collaboration | www.campbellcollaboration.org

11 Search strategies

11.1 PUBMED: 115

(((nurse) AND (practitioner* or specialist* OR clinician*)) OR "Nurse Clinicians"[Mesh] OR "Nurse Practitioners"[Mesh]) AND (medical practitioner* OR general practitioner* OR medical professional OR medical professions OR doctor OR doctors OR physician* OR "Physicians"[Mesh]) AND (rape OR rapes OR molest* OR ((sex OR sexual OR child OR children OR adolescent OR adolescents) AND (assault OR offence OR offences OR offense OR offenses OR unlawful sexual intercourse OR abuse OR abuses OR offence OR offences OR crime OR crimes OR delinquency)) OR "Sex Offenses"[Mesh])

11.2 EMBASE: 90

1. (nurse and (practitioner* or specialist* or clinician*)).af. 2. exp nurse practitioner/ 3. 1 or 2

4. (medical practitioner* or general practitioner* or medical professional or medical professions or doctor or doctors or physician*).af. 5. exp physician/ 6. 4 or 5 7. (rape or rapes or molest* or ((sex or sexual or child or children or adolescent or adolescents) and (assault or offence or offences or offense or offenses or unlawful sexual intercourse or abuse or abuses or offence or offences or crime or crimes or delinquency))).af.8. exp sexual crime/ 9. 7 or 8 10. 3 and 6 and 9

11.3 COCHRANE (ISSUE 2): 9

#1 (nurse and (practitioner* or specialist* or clinician*)) #2 MeSH descriptor Nurse Clinicians explode all trees #3 MeSH descriptor Nurse Practitioners explode all trees

53 The Campbell Collaboration | www.campbellcollaboration.org

#4 (#1 OR #2 OR #3)#5 (medical practitioner* or general practitioner* or medical professional or medical professions or doctor or doctors or physician*)#6 MeSH descriptor Physicians explode all trees#7 (#5 OR #6)#8 (rape or rapes or molest* or ((sex or sexual or child or children or adolescent or adolescents) and (assault or offence or offences or offense or offenses or unlawful sexual intercourse or abuse or abuses or offence or offences or crime or crimes or delinquency)))#9 MeSH descriptor Sex Offenses explode all trees#10 (#8 OR #9)#11 (#4 AND #7 AND #10)

11.4 SCI: 65

# 1 TS=(nurse and (practitioner* or specialist* or clinician*)) # 2 TS=(medical practitioner* or general practitioner* or medical professional or medical professions or doctor or doctors or physician*) # 3 TS=(rape or rapes or molest* or ((sex or sexual or child or children or adolescent or adolescents) and (assault or offence or offences or offense or offenses or unlawful sexual intercourse or abuse or abuses or offence or offences or crime or crimes or delinquency))) # 4 #3 AND #2 AND #1

11.5 GOOGLE SCHOLAR: 12

split into 2 searches because of word limit. Results 2 + 12 (the 2 results are included in the 12 results; total =12) none relevant

Nurse; (practitioner or specialist or clinician); (medical practitioner or general practitioner or medical professional or medical professions or doctor or doctors or physician); (rape or rapes or molest* or assault or offence or offense)

Nurse; (practitioner or specialist or clinician); (medical practitioner or general practitioner or medical professional or medical professions or doctor or doctors or physician); (unlawful sexual intercourse or abuse or crime or delinquency)

54 The Campbell Collaboration | www.campbellcollaboration.org

Additional searches 26/02/2012

The following three searches were conducted on the following databases: AMED BNI Cinahl Health Business Elite HMIC PsycInfo Social Policy & Practice

1. Sexual AND assault AND nurse AND examiner

1. exp Violence/ or exp Physical Examination/ or exp Forensic Nursing/ or exp Rape/ 2. exp Specialties, Nursing/ or "Coroners and Medical Examiners"/ or exp Forensic Medicine/ or exp Rape/ 3. exp "Reproducibility of Results"/ or exp Forensic Medicine/ or exp Sex Offenses/ or exp Crime Complainants/ or sexual offence examiner.mp. or exp Physical Examination/ 4. (quality or reliability or credibility).mp. [mp=protocol supplementary concept, rare disease supplementary concept, title, original title, abstract, name of substance word, subject heading word, unique identifier] 5. 1 and 4 6. 2 and 4 7. 3 and 4 8. 5 and 6 9. 5 and 7 10. 6 and 7 11. remove duplicates from 8 12. remove duplicates from 10 13. conviction.mp. 14. 9 and 13 15. 10 and 13 16. 8 and 13 17. prosecut*.mp. [mp=protocol supplementary concept, rare disease supplementary concept, title, original title, abstract, name of substance word, subject heading word, unique identifier] 18. 9 and 17 1. Sex offenses/ or Sexual abuse/ or rape.mp. or Jurisprudence/ 2. sexual assault.mp. 3. molest*.mp. [mp=abstract, heading words, title] 4. Violence/ or Crime/ or "Predictive value of tests"/ or sexual crime*.mp. 5. unlawful sexual intercourse.mp. or Legislation/

55 The Campbell Collaboration | www.campbellcollaboration.org

6. 1 or 2 or 3 or 4 or 5 7. Knowledge/ or "Delivery of health care"/ or Nursing care/ or Nurses/ or Complainant care/ or nurse practitioner.mp. or Jurisprudence/ 8. 6 and 7 9. doctor*.mp. 10. physician*.mp. or Physicians/ or Jurisprudence/ 11. Professional practice/ or "Cost benefit analysis"/ or general practitioner*.mp. or Methods/ 12. medical profession*.mp. or Professional practice/ 13. 9 or 10 or 11 or 12 14. 8 and 13

56 The Campbell Collaboration | www.campbellcollaboration.org