Prisoner-on-prisoner drug searches in prisons in England and Wales: ‘Business as usual’

Prisoner reported drug and contraband searches in adult men’s prisons in England and Wales represented almost a quarter of reported and recorded ‘sexual assaults’ from 2004 to 2014. These searches are more likely to involve multiple perpetrators and weapon use than other types of sexual assaults and are most frequently carried out in the relative privacy of a cell. The research presented here is based on an analysis of Her Majesty’s Prison and Probation Service (formerly the National Offender Management Service) Incident Recording System data, providing insights into the proportion of recorded sexual assaults which are related to drug searches. This analysis enables a distinction to be made between prisoner-on-prisoner drug and contraband searches and other sexual assaults. Analysis shows that prisoner-on-prisoner searches are frequent, often pre-meditated, brutal and appear to be an accepted aspect of everyday prison life.

prisons has not been available to researchers. In contrast, international research, mostly from the United States, has sought to elucidate levels of sexual abuse in prisons and the characteristics of incidents, even passing legislation to require prisons to adhere to reporting and prevention standards (Gaes and Goldberg, 2004;Struckman-Johnson and Struckman-Johnson, 2013). Some formative efforts, such as those based on health improvement agendas (Green et al., 2003) or prisoners' rights (Stevens, 2015), have sought to ascertain staff and prisoners' experiences and views of prison-based sex and sexual assaults. The Howard League for Penal Reform, a UK-based, campaigning charity whose mission is to work towards less crime, safer communities and fewer people in prison has also sought access to prison records and other avenues of data to explore the state's response to sexual assault among those incarcerated in England and Wales.
In 2015, following a protracted application process, the National Offender Management Service (NOMS) provided to one of the authors a full administrative data set including all reported sexual assault cases for the period 2004-2014. The data presented an opportunity to provide a descriptive, statistical overview of characteristics of incidents and those involved in them. It also allowed for an exploration of the reporting process itself. Initially, the broad research question was confined to identifying initial responses to allegations of reports of prisoner-on-prisoner sexual assault. It was anticipated that as well as revealing patterns and characteristics of these assaults, the data would provide insights into the prison officer-generated incident descriptions (IDs) about the initial response to prisoner-on-prisoner sexual assaults. The research did indeed reveal patterns, trends and characteristics and a number of insights into incidents and the responses to those incidents in the period under review. One of those insights was the prevalence of 'sexual assaults' that were clearly linked to intimate drug and contraband searches.
This article focuses on those sexual assaults which came to be coded as prisoner-on-prisoner drug or contraband searches.

The sample
The full sample of 844 adult male prisoner-on-prisoner sexual assaults involved a total of 2032 prisoners 1 in incidents which occurred between 2004/2005 and 2013/2014 in prisons in England and Wales. Of these, 1116 prisoners were recorded as perpetrators, 826 as victims and 90 as 'other', that is, present during the reported incident, noted as a witnesses or with an undetermined role in the reported incident. Following detailed analysis, only 583 of the 844 sexual assaults contained sufficient information to allow for the precise nature of the sexual assault to be determined. Of those 583 incidents, just under a quarter could be identified as being related to a prisoner-on-prisoner drug or contraband search and the remainder involved other behaviour or offending from verbal threats to touching and rape.
The sample excludes incidents involving staff and those where anyone involved -in any capacity -was under 18 years of age. 2 The sample included all prisoner-on-prisoner incidents involving adult males, substantiated or unsubstantiated, classified as being a 'sexual assault' and recorded on the Incident Recording System (IRS) and managed during the data sample period by NOMS. These incidents only represent those identified, reported, classified and recorded by prison staff and it should be assumed that they do not represent the full extent of prisoner-on-prisoner sexual assaults which took place during the data period. Each incident involves one or more individuals.
These findings are the only existing published analysis of the patterns and initial responses to reported sexual assault incidents in England and Wales based on the officially recorded data. The data sets have been developed from the previously untapped IRS and trends suggest that reported sexual assaults in prisons take many forms and are, with some minor fluctuations, increasing. 3 For the purposes of this article, prisoner-on-prisoner drug or contraband searches include incidents where the report makes reference to an intimate internal search using objects or digits or an intimate but external inspection of the genital area for the purpose of finding smuggled goods such as drugs or mobile phones.

Quality of the recording data
Upon receipt, the raw IRS data were disorganised and of poor quality, perhaps explaining why there had been no concerted attempt to analyse it previously. Extensive data cleansing, reorganisation and coding led to the development of two new data sets, which together provide a strong foundation for future research. Analysis of the qualitative data fields identified, where sufficient information was recorded, a more precise understanding of the nature of sexual assaults, which are recorded in the IRS simply as 'sexual assaults'. Recording processes vary across the prison estate; however, the IRS reveals that most sexual assaults are reported verbally by victims directly to a member of staff, most often a prison officer. Details of reports of sexual assaults are recorded in the Wing observation book, written up as an intelligence report and escalated to custodial managers who supervise the prison response and any referrals made to police. Incident reports are generally entered into the IRS by custodial managers, based on the information given to them by the prison officer accounts. Most fields in the IRS require quantitative information with the exception of the single qualitative field in which a narrative description of the incident is included.
Drug and contraband searches involved brutality, violence, restraint, multiple perpetrators and everyday objects adapted for use as weapons. These 'incidents' often took place in the prisoner's own cell or another cell, away from the gaze of closed-circuit television (CCTV) and surveillance by staff. This article provides explicit and at times, disturbing descriptions of drug and other contraband searches derived from the analysis of reports recorded in prisons across England and Wales in the period 2004-2014. It indicates that these searches form part of normal life in prisons in England and Wales and that incident patterns suggest that rather than being subject to routine prevention measures, they have become 'business as usual'.
Research about drugs, associated markets and supply routes in prisons is sparse (Crewe, 2006: 348), although evidence about levels of consumption are well-documented (Her Majesty's Inspectorate of Prisons (HMIP), 2014(HMIP), , 2015Hopkins and Brunton-Smith, 2014). Equally neglected has been the reporting, recording and response to drug-related sexual assaults in prisons in England and Wales. Findings from this study suggest that prisoner-on-prisoner sexual assaults are mostly reported directly by victims to prison officers.

What we know
Research-based estimates of the incidence and prevalence of sexual victimisation in prisons have been dominated by studies in the United States (Beck and Johnson, 2012;Gaes and Goldberg, 2004;National Institute of Justice, 2006). Australian studies have provided some analysis of incidence levels (Butler et al., 2002;Simpson et al., 2016;Steels and Goulding, 2009;Yap et al., 2011), with some early, small-scale studies from other countries (Einat, 2013;Gear, 2007;Papadakaki et al., 2019).
U.S. researchers have focused on prison officer perceptions of sexual behaviour, homosexuality, risk, race, age and sexual victimisation Eigenberg, 2000aEigenberg, , 2000bHensley and Tewksbury, 2005;Moster and Jeglic, 2009) providing strong comparative value for researchers. While there is a wealth of U.S. research-based information on prevalence and observational studies of sexual behaviour in men's prisons, research from the United Kingdom generally is limited. The research that does exist is constrained either by being small-scale or overreliant on ex-prisoners' accounts; there being very little published comparative research about the perceptions and responses of prison officers and police to sexual victimisation (see Banbury, 2004;Banbury et al., 2016;O'Donnell, 2004;Stevens, 2015Stevens, , 2017. Early U.S. research on victimisation in prisons concentrated on sexual roles -consensual and coercive -adopted by prisoners (Bowker, 1980;Donaldson, 2001;Lockwood, 1980). Other research (e.g. Clemmer, 1940;Eigenberg, 1992;Sykes, 1958) describes the sexual and violent behaviour of prisoners, with researchers often dividing their 'personas' into the 'predators' and the 'victimised' and the 'masculine' and the 'feminine'. The ambiguity and fluidity of prison sexuality and sexual assaults also features strongly in the literature (e.g. Gear, 2007;Groth, 1979). Studies which focus on sexuality in U.S. prisons and include a specific emphasis on finding out about the behaviour and attitudes of, and towards, gay and bisexual men (in protective custody settings) have provided some insights into the impact of incarceration on sexuality (Bowker, 1980;Donaldson, 2001;Lockwood, 1980). Prison staff may also experience problems in determining the exact nature of some sexual behaviour in prisons (Eigenberg, 2000a(Eigenberg, , 2000bHensley and Tewksbury, 2005;Stevens, 2013: 1), particularly in identifying coercive 'relationships'. Issues faced by prison officers in determining the existence and nature of any sexual activity, consensual or coercive, have been noted in the United Kingdom by the Prisons and Probation Ombudsman 4 (2013: 5). The problem for prison staff in successfully distinguishing between consensual sex and coercive relationships was acknowledged by Stevens (2013: 3) who noted that prison governors had expressed concerns about whether sex could be truly consensual in prison environments where rational choices were restricted. Related literature on sexuality and masculinity are the focus of a number of studies of behaviour in men's prisons, again in the United States (Bowker, 1980;Lockwood, 1980;Sabo et al., 2001;Saum et al., 1995;Toch, 1998;Wooden and Parker, 1982) but are less represented in the U.K. context (Maycock and Hunt, 2018;Schinkel, 2014;Sloan, 2016).
The prevalence of drugs in prison (Crewe, 2009), an underlying drug culture (Jewkes, 2005;Kupers, 2001), levels of consumption (Hopkins and Brunton-Smith, 2014;Penfold et al., 2005) and more recently, Ralphs et al.'s (2017) work on the use of synthetic drugs by prisoners have all contributed to a more nuanced understanding of drug use in U.K. prisons. However, until Banbury et al.'s (2016) study, which developed the first typology to include drug searches as a dimension of sexual assaults, this aspect had been largely absent from the research. The lack of coverage of drugrelated sexual assaults was noted by Banbury (2004: 123) as a significant research gap because prisoner-on-prisoner-forced drug searches had failed to be identified as an issue, despite being the most frequently reported incidents by her sample of male and female ex-offenders. Banbury (2004: 124) stated that drug-related sexual coercion provoked fear and presented additional barriers to reporting to the prison authorities because of fear of forced abstinence and concern about disciplinary action for involvement in drug-taking. Subsequent work (Banbury et al., 2016: 383) identified that drug searches were more likely than other types of sexual assault to involve more than one perpetrator. Banbury et al. (2016: 377) developed a typology of prisoners involved in sexual assaults which included distinct types of drug searchers: one involved orchestration of the search, where other prisoners were intimidated into carrying out the search on behalf of a 'watcher', who, besides being a lookout, reported sexual gratification. Banbury et al.'s (2016: 377) second type of drug search perpetrator used a hand or implement inserted into the victim's anus but reported no sexual motive.
Perpetrators carrying out these drug searches described compassion for their victims and saw their victimisation as a means to an end, often avoiding intimate physical contact by using objects and lubrication to accelerate the removal of drugs or contraband (Banbury et al., 2016: 378) In contrast, U.S. prevalence studies have generally focused on rape and serious sexual assault (Gaes and Goldberg, 2004), mostly avoiding 'less serious' incidents and prisoner-on-prisoner drug and contraband searches as a form of sexual assault.

Incident Recording System
The IRS is an administrative database introduced in the 1980s by the U.K. government for the purpose of recording and management of information relating to deaths in custody, self-harm and assaults. Sexual assaults are included in the IRS as a sub-set of serious assaults. The IRS includes fields about various dimensions of the recorded sexual assault, including demographic characteristics of those involved and a text-based field for qualitative IDs, which are completed by prison staff and often provide additional information about the assault and response activities carried out by staff. The IRS is the source of the published official statistics in the form of Ministry of Justice Safety in Custody Bulletins. These Bulletins, first published in 2010, report the trends of deaths in custody, self-harm and assaults (on staff and prisoners) in prisons in England and Wales. In 2018, the Ministry of Justice (MOJ) (2018: 5) indicated that the introduction of the IRS and improvements in centrally held data now meant that there were consistent data for assaults (from 2002) from which to determine trends. The Safety in Custody Bulletins include sexual assault incidents as a sub-set of serious assaults. Serious assaults are defined as: . . . those which fall into one or more of the following categories: a sexual assault, requires detention in an outside hospital as an in-patient; requires medical treatment for concussion or internal injuries; or incurs any of the following injuries: a fracture, scald or burn, stabbing, crushing, extensive or multiple bruising, black eye, broken nose, lost or broken tooth, cuts requiring suturing, bites, temporary or permanent blindness. (MOJ, 2020: 6) In 2015, the MOJ published an extraordinary management Bulletin on sexual assaults that provided some limited trend analysis of incidents reported in the period 2002-2013. It included a breakdown of the proportion of incidents occurring on each day of the week, the proportion resulting in minor and serious injuries and the number of referrals to police and incidents subject to adjudication (MOJ, 2015). This MOJ Bulletin was described as being an 'ad hoc' analysis of sexual assaults in prisons because the findings that it contained could not be included as part of the Safety in Custody Bulletins because the data did not meet the quality thresholds to be designated as 'National Statistics'. In particular, the recorded outcomes of sexual assault were not deemed to be of sufficient quality to be classified as national statistics (MOJ, 2015). The Bulletin noted that the category of 'sexual assaults' covered a wide range of incidents: . . . from rape to inappropriate touching. Reported sexual assaults will also include incidents where there are attempts to retrieve drugs or other prohibited items that may be hidden on the victim. Whether an incident is deemed to be a sexual assault is determined by the perception of the victim. Any assault incident reported as a sexual assault is classified as a serious assault. (MOJ, 2015b: 2) In practice, this broad definition of 'sexual assault', as determined by the perception of the victim, may hold benefits in terms of encouraging reporting. However, the approach adopted to record all incidents simply as 'sexual assaults' results in a lack specificity in the IRS data and presents a barrier for the prison service in identifying more precise information about the seriousness and nature of sexual assaults in prisons. Importantly, prisoner-on-prisoner drug and contraband searches have been indistinguishable from other sexual assaults.
Analysis of the IRS as an administrative database was identified as the most appropriate method to develop new data sets to provide a baseline of descriptive statistics as well as identifying reporting and recording practices associated with prisoner-on-prisoner sexual assaults. The multiplicity of practical, methodological and ethical issues associated with conducting research in (and on) prisons have been well-documented by sociologists, medical researchers and psychologists (Day et al., 2005;Rivlin et al., 2012;Rubin, 1976;Sykes, 1958). Administrative data analysis was preferred as a method, not only to protect vulnerable populations from further harm (Halliday, 2019;Hewson, 2017), but also to develop the first baseline analysis of recorded sexual assaults in men's prisons in England and Wales. Hitherto, the available published data have remained restricted to the trends published by the MOJ as Safety in Custody Statistics.
Following extensive discussions and applications to NOMS, a series of data files were provided in October 2014. The files included all reported and recorded prisoner-on-prisoner incidents between 2004 and 2014 from prisons in England and Wales which involved adult male prisoners, aged 18 or over. Overtime, the data were sifted and organised into two new data sets, both of which provided insights into the level and nature of reported and recorded sexual assaults. The first data set developed an assessment of the individual and institutional characteristics of reporting and recording of sexual offences in U.K. adult male prisons. It allowed for descriptive statistical analysis of 'incidents' (N ¼ 844) and 'involvements' (N ¼ 2032; a grouping which includes victims, perpetrators and 'others').
The second data set comprised an analysis of qualitative data fields or prison officer-generated IDs from incident-level data (N ¼ 844). The incidents were analysed using NVivo 10.0 software to identify themes and create more detailed information about the reported sexual assaults. The selection of NVivo 10.0 as a tool for organising and coding the IRS data was based upon the requirement for a suitable method to manage the large data set in a secure environment. The research adopted a pragmatic coding strategy, using a process by which the data were decoded (to understand its meaning) and then encoded to develop labels and categories therefore, combining different types of codes for analysis of the data. Initial coding of 50 IDs identified that the most appropriate method of coding the remaining IDs was to code the entire ID with multiple codes, rather than coding the entire ID with a single code. Use of this technique, known as 'splitting', provided greater insight into the IDs, by highlighting nuances and creating a more detailed analysis of each incident. This coding approach allowed for the quantification of the qualitative field (ID), where sufficient detail was provided, and devised more precise identification of the type of sexual assault in 583 of the 844 cases.

Prisoner characteristics associated with involvement in drug searches
The detailed sexual assault types covered a broad spectrum of occurrences from verbal sexual threats (18) to actual or attempted anal rape with a penis (106). The most frequent sexual assault recorded, related to genital touching (138), followed by prisoner-on-prisoner drug searches (136), actual or attempted anal rape with a penis (106) 5 and unwanted physical contact (82) (see Figure 1). In her study of ex-offenders from British prisons, Banbury (2004: 123) found that threats of 'coercive sexual behaviour' and 'forced drug searches' were the most commonly experienced sexual assaults. Supporting Banbury's (2004) findings, data set 1 confirmed that prisoner-onprisoner drug and contraband searches are more commonly reported, and perhaps experienced, by prisoners than rape -which remains relatively rare in adult men's prisons in England and Wales. The IRS data do not allow for any analysis of the motivation for carrying out prisoner-onprisoner drug and contraband searches. While it can be assumed that they are largely carried out for economic purposes, sexual motivation, as identified by Banbury et al. (2016), cannot be ruled out. Figure 1 shows that of the sexual assaults where the type of incident could be identified (583), just under a quarter involved a prisoner-on-prisoner drug search. This number is likely to be an underestimate because of the risks involved in reporting and being seen by other prisoners to be 'snitching' as well as the perceived need to conceal drug dealing from staff. Other incidents coded as 'digital penetration of the anal passage' and 'anal penetration with an object' were also likely to be drug or contraband searches. These incidents were not included in the analysis of drug searches because strict coding rules were applied and where there was no explicit reference to drugs in the ID, incidents were not classified as such. Further cases coded as 'touching genitals' or 'inspecting genitals' were also likely to be associated with prisoner-on-prisoner drug or contraband searches but similarly were not counted. Prisoners' reluctance to admit that incidents were associated with drug use and smuggling in prison drug markets has potentially reduced the number of IDs referring to drug searches because prisoners only made limited disclosures to prison staff.
The phenomena of prison drugs markets and their impact on the dynamics of prison life have been under-researched (Crewe, 2006: 348). A partial explanation for the lack of scrutiny of drug supply markets in U.K. prisons is that research has identified that a common route is via staff, smuggling for financial benefit (Crewe, 2006: 356;Ralphs et al., 2017: 62). However, failures to reduce drug smuggling and use may be based on the pervasive importance of drugs to everyday prison life in the United Kingdom. During his study of an English prison, Crewe (2005: 161) noted that 'drug taking and dealing were accepted by officers and prisoners as inevitable, almost banal, features of the inmate world'.

Institutional characteristics associated with drug searches
Fluctuation across the prison estate in England and Wales over the 10-year reporting period (e.g. in security classification, capacity, type, role and function) prevented analysis of any differences in reporting and recording between private and public prisons. However, prisoner categorisations were used to examine patterns of involvement in sexual assaults. The proportion of categories 6 A and B prisoners involved in all sexual assaults are broadly proportionate to their make-up in the general prison population. 7 However, prisoners with lower security categories (C and D) or those unsentenced (convicted but awaiting sentencing in court) were under-represented. Forty-five percent of prisoners involved in a sexual assault were category C (this was true for both perpetrators and victims), 1% were category D and 3% were unsentenced compared with 57% (category C), 9% (category D) and 12% unsentenced of the general male prisoner population. More notable is the over-representation of unclassified (convicted and sentenced) prisoners: a third of all prisoners involved in a sexual assault were awaiting classification, compared to only 1% of the male prison population. The over-representation of unclassified prisoners is linked to risks associated with the early or 'Reception' stage of being housed at a prison. Staff in category B local prisons take responsibility for the initial classification of prisoners, which then determines where their sentence will be carried out. While prisoners classified as category C were under-represented in the reporting data, many were located in category B local prisons at the time of the reported sexual assault. This was also true of unclassified prisoners who were being held in category B local prisons while awaiting classification. Analysis of the prisoner's offence for which the sentence was being served indicated that perpetrators of prisoner-on-prisoner drug and contraband searches were most likely to have been convicted for robbery or burglary. In contrast, perpetrators of other sexual assaults (non-drug searches), ranging from verbal threats to rape, were more likely to be in prison for violence against the person or sexual offences.
Thirty-eight percent of drug or contraband searches occurred less than a week from first arriving at a prison. The finding of a clear risk of being subject to a prisoner-on-prisoner drug or contraband search shortly after reception at a prison is supported by Ralphs et al.'s (2017: 62) research in an English category B local prison. Ralphs et al. (2017) note that prisoners often used their recall to prison for drug supply purposes, 'plugging' drugs into their anus as a commercial venture. It is not possible to confidently determine the frequency of prisoner-on-prisoner drug and contraband searches specifically linked to licence recalls from the data because the likelihood is that a proportion of these were probably never reported as sexual assaults. Similarly, the IRS data may also exclude incidents concerning professional drug suppliers where smugglers enjoy the protection of gangs (and their customers) and therefore, may not be subject to 'screening' or prisoner-onprisoner searches because they import goods to order or specification. In these circumstances, recalled prisoners who are 'plugged' would not be forcibly searched by other prisoners because organised internal suppliers expect them to willingly defecate drugs or contraband to supply illegal drugs markets. Prisoner-on-prisoner screening possibly takes place as a routinised check on new inmates or to interrupt potential supply by stealing concealed drugs and contraband.
Beyond risks during the first week following reception, two thirds of the prisoner-on-prisoner drug and contraband searches occurred in the first month, compared to 28% of non-drug searchrelated sexual assaults. In both prisoner-on-prisoner drug and contraband searches and other sexual assaults, victims and perpetrators may also have been at an increased risk of involvement because of their individual characteristics.
Black, Asian and minority ethnic (BAME) perpetrators were over-represented in prisoner-onprisoner drug or contraband searches; 41% of perpetrators in drug and contraband searches identified as being Black, Asian or minority ethnic when they only accounted for 25% of the overall sexual assault perpetrator population and general prison population. Conversely, prisoners identified as being White (75% of the prison population) were under-represented as being involved in drug and contraband searches. White men made up 73% of all sexual assault perpetrators but only 5% of those involved in prisoner-on-prisoner drug or contraband searches and hence were more likely to be involved in other types of sexual assault. It was not possible to accurately analyse ethnic background by role in prisoner-on-prisoner drug searches. In all other types of sexual assault (e.g. verbal threats to anal rape), the ethnic background of the perpetrators was broadly in line with the overall perpetrator population.
Younger age groups were over-represented in all sexual assaults, with two thirds of victims (66%) and over half of perpetrators (55%) being under 30 at the time of the assault. This age group accounted for just under half of the average general prison population throughout the period. There was no difference between the proportion of those aged under 30 who were victims of prisoner-onprisoner drug and contraband searches (62%) or other sexual assaults (65%).

Multiple perpetrators
Incidents involving prisoner-on-prisoner drug and contraband searches were 8 times more likely to involve multiple perpetrators than other types of sexual assaults. Where the number of perpetrators was recorded, 78% of drug and contraband searches involved multiple perpetrators compared to 10% of all other sexual assaults. The number of people involved in a single incident ranged from one to nine individuals. Nearly two in three (N ¼ 531) of the reported incidents involved two people -one perpetrator and one victim. Figure 2 shows the proportion of incidents where multiple perpetrators were recorded as being involved. Data set 2 clearly showed that drug and contraband searches were more likely than other types of sexual assault to involve multiple perpetrators. IDs linked to multiple perpetrator incidents included references to anal drug searches using digits (28) as well as improvised enemas (10) which involved use of water or shower gel bottles with soapy water and tubes, sometimes created from biros. The following is an example of a targeted contraband search where the victim identified five named perpetrators performing a search for a mobile phone: Prisoner V reported to unit staff that he had been assaulted the previous evening during final lock up at approximately 20.00 hours. He alleged that he had been sexually assaulted in cell 115 on Delta wing, that of Prisoner A. He stated that offenders Prisoner B, Prisoner C, Prisoner D, Prisoner E and Prisoner A had all been present when he . . . had been forced in to the cell, had his trousers and boxer shorts pulled down so that his lower half was naked and that offenders listed above had then searched under his testicles by moving them to one side and had also pulled apart his buttock cheeks to search his anal area. He stated that they told him they were looking to take possession of a mobile telephone they believed he had. Subsequent examination of CCTV evidence corroborated this and showed that Prisoner D had kept watch outside the cell. The offenders were relocated to the segregation unit pending adjudication for assault.
Analysis of the IDs shows that in these types of searches, perpetrators are often named by the victim and are therefore known to them. These multiple perpetrator drug and contraband searches frequently include forcible stripping and an external visual and physical inspection or internal intimate search using objects.
Where IDs did include information about the mechanisms used to carry out multiple perpetrator sexual assaults (145), victims were more likely to have been restrained by other prisoners, threatened or overcome with violence than in other non-drug-related sexual assaults.
Based on the IDs, the most common type of drug search was anal penetration with an object (40). This 2010 ID represents the brutal nature of these searches: It is reported that Prisoner A assaults Prisoner V by pulling his trousers down and held him round the throat in an attempt to get secreted drugs. [V]ictim and perp plans opened, incident report completed . . . Following this incident Prisoner V stated he was violated by Prisoner A who he alleges inserted his finger into his anus and also tried to insert a fork in a bid to remove drugs while Prisoner B held him by the throat. Prisoner C was also present whilst this was taken place. Police have been informed even though Prisoner V did not want police involvement.
A central difficulty for prisons is that prisoner-on-prisoner drug and contraband searches are some of the most intrusive and violent sexual assaults which epitomise rule breaking but seem to form part of everyday life in many prisons. Although relatively few reports of drug and contraband searches are recorded across the prison estate, Ralphs et al.'s (2017) study and the 2013-2014 annual report of HMIP (2014) highlighted the increased use of new synthetic drugs in prisons. Ralphs et al. (2017) asserted that new prison drug markets, which emerged around 2014, presented high risks of harm for prisoners and staff because of unprecedented access to, and use of, synthetic drugs. Figure 3 shows, in line with Ralph's (2017) assertions, that reporting levels for drug and contraband searches increased steadily over the data capture period until 2014 when reported drug and contraband searches almost tripled. The IDs showed similar patterns of staff routine activity in their response to reports, often caring and professional but mostly performed without any real prospect of a criminal justice-based outcome. In terms of consequences, there were no differences in the data between outcomes for prisoner-on-prisoner drug and contraband searches and other sexual assaults. Figure 4 shows that the most commonly recorded 'outcome' was being 'subject to a police investigation'. In practice, the IRS recording data gave little insight into the differences between being 'subject to a police investigation' and being 'referred to the police'. Inconsistences in the quality of recording of outcomes have been outlined earlier in the article as a key reason for 'outcomes' not reaching the quality standards to be classified as official or national statistics.
The IDs suggested that prisoners regularly made reports without the prospect or intention of realising criminal justice outcomes, but perhaps as a short-term means to an end for attaining access to medical attention or segregation from perpetrators. The ID below describes a prisoneron-prisoner drug search following a prison move, in which multiple perpetrators targeted the victim. The prison officer noticed the prisoner's facial injuries and prompted the victim to disclose the sexual assault. Although vigilant, this officer did not record any realised outcomes, only noting an internal prison investigation, recorded solely for information: At lock up after association it was noted that Prisoner V had received facial injuries.
[He] stated that he had been grabbed by several unknown . . . prisoners beaten with punches and kicks and held down whilst his anal cavity was searched with a spoon. V returned to Prison X on X/X/07 from Prison Y and was rumoured to be in the possession of drugs and a mobile phone anally secreted. Prisoner V refused to cooperate as to his assailants and did not want the matter referred to the police but went to HCC [healthcare]. The assault was believed to have occurred in cell GX3-0X. A further ID from 2007 recorded two separate multiple perpetrator sexual assaults, use of a weapon, physical assault and pre-meditation without any mention of the category C victim's condition/injuries, staff responses or an outcome. The outcome code relating to this incident from the IRS was recorded as being 'subject to a police investigation'. However, because the ID was not updated following the initial response, without having access to police records, the final result of this incident remains unknown to the researchers and the prison service: Prisoner V has alleged that he was assaulted by Prisoner A. V claims that Prisoner A produced a metal implement to force him onto the floor. It is then alleged that Prisoner A forced V to drop his trousers and undergarment put on a latex glove and insert his fingers into A's anus in an attempt to retrieve 3 balls of heroin. Prisoner V was then allegedly assaulted by Prisoner B who pinned V against a wall and forced him to handover a further 3 balls of Heroin. Shortly after this V alleges that he was dragged into another cell by Prisoner C and forced to drop his trousers and bend over but there was no more heroin to be found.
Multiple perpetrator incidents were much more likely to involve use of a weapon (40%) than those with a single known perpetrator (6%).
In most recorded sexual assaults, no weapon was used (85%). A weapon was reported in 120 incidents, the most frequently reported was a spoon (N ¼ 49), followed by a bottle (N ¼ 13) and a knife (N ¼ 12). Other weapons included sharp objects such as razors, unspecified blades, forks, blunt weapons including brush handles, toothbrushes, toilet brushes, yoghurt pots and pens.
An ID from 2009 provides some typical context about prisoner-on-prisoner drug and contraband searches. It also references multiple perpetrators, use of weapons and took place in a cell resulting in the victim visiting healthcare and then being moved to segregation for his own protection. The ID identified that weapons were used both as a mechanism to threaten and overcome the victim and as tools to carry out an internal anal search: Outcome not recorded Prisoners A, B and C assaulted Prisoner V in his cell. He was held in his cell for 1 hour 30 minutes and was threatened with a knife and a screwdriver and his anus was invaded with a plastic bottle and a fork. Prisoner A had the screwdriver and put a glove on and placed a finger up his bottom to see if he had a mobile phone plugged. Prisoner B held a 3/4'' knife to his neck. Prisoner V sustained bruising to leg and back, had scratches to face, seen by healthcare and relocated to seg for own protection.
It is apparent from the IDs in this article that where a location is mentioned, sexual assaults tended to take place in the relative seclusion of a cell, which would not be subject to CCTV monitoring. So, while it may be true that shared cells offer opportunities for solidarity with cell mates (Jewkes, 2005) and may provide safety and relief from the demands of prison life (Crewe et al., 2013;Sloan, 2016), they are also a location for brutal assaults and prisoner-on-prisoner drug and contraband searches.

Conclusion
This article contributes significantly to what we know about prisoner-on-prisoner-reported drug and contraband searches in prisons in England and Wales. Drawing on the data provided by NOMS for the period 2004-2014, the analysis demonstrates that these searches were reported more frequently than rape, representing almost a quarter of reported and recorded 'sexual assaults' in the overall data sets. Such searches, frequently pre-meditated, often involved brutality, violence, restraint and everyday objects adapted for use as makeshift weapons. These searches presented challenges for coding both in terms of the actual task of coding and in the knowledge that each case represented the victimisation of an individual.
Analysis showed that drug and contraband searches are more likely to involve multiple perpetrators who were identified as being Black or Asian minority ethnic. As victims, BAME prisoners are under-represented in the data, perhaps indicating that they are either less willing to report sexual assaults or that they are less likely to be victimised. Prisoner-on-prisoner drug and contraband searches are most frequently carried out in the relative privacy of the 'safe' cell. The victims of these searches are often new to the prison, frequently targeted in the first 3 months of entering an establishment while awaiting classification. Reception stage-related risks were clear for prisoners subject to prisoner-on-prisoner drug or contraband searches because 38% of these were reported as occurring less than a week from entering the prison.
IDs were included here. We accept that these examples were graphic and explicit. However, the decision was made to include them verbatim rather than clean them up for easier consumption. We are in agreement with Davis (1968: 9) from his early study of prison sexual assaults in Philadelphia, where he resolved to quote incidents verbatim because 'the incidents [were] raw and ugly. Any attempt to prettify them would be hypocrisy'. Analysis shows that prisoner-on-prisoner drug and contraband searches are frequent, often pre-meditated and are a brutal and accepted aspect of everyday prison life.
Statistics alone can only provide information as it has been reported and recorded. The findings associated with this research highlight the limitations and gaps in the IRS data. These limitations have the potential to cast doubt on the value of statistical bulletins which focus solely on the extent of sexual assaults in prison. The IRS data, provided by NOMS, obscure the scale and nature of the problem of prisoner-on-prisoner drug and contraband searches and this research emphasises the problem of not differentiating between types of sexual assault in the recording processes. If the MOJ wishes to address the ambiguities of incidents recorded as 'sexual assault' in prisons, a first step would be to make a clear distinction in its recording practices between prisoner-on-prisoner drug and contraband searches and other sexual assaults.