Scottish Government

CHAPTER THREE REVIEW OF LITERATURE

Introduction

3.1 This review of the literature is divided into three main sections. The first section will look at emergency responses, the second focuses on mainstream and emerging interventions with the final section looking at a range of early individual and social indicators facing those most at risk of drug overdose.

3.2 In section 1, the reviewed literature on emergency responses covers witness response and intervention to drug overdose. It also examines ambulance, police and accident emergency responses. The section concludes by looking at the role of public health alerts.

3.3 In section 2, the key themes explored are traditional mainstream treatment interventions. This involves examining the role of methadone, General Practitioners and engagement and retention themes for drug users and service providers. This section concludes by exploring 2 emerging interventions: Take-Home-Naloxone and Safer Injecting Rooms.

3.4 Finally, in section 3 early individual and social indicators facing those most at risk of non-fatal and fatal overdose are examined. The indicator themes are drug users' tolerance, the impact of injecting drug use and poly drug use. The health-related indicators include morbidity, recent life problems, the role of antidepressants and suicide. Social indicators explored are the role of drug using networks, public injecting and accommodation issues.

Section 1 - Emergency Responses

Overdose Witnesses and Interventions

3.5 Witnesses present at an overdose event are willing to intervene, according to the papers reviewed. For example, Best et al (2002) noted that witnesses reported using a range of strategies - from appropriate (e.g. cardio pulmonary resuscitation) to inappropriate (e.g. shocking the casualty with cold water). The authors noted that successful outcomes were strongly linked with immediate overdose onset while fatalities were often linked to slow overdose onset. It has also been suggested that where narcosis is slow to develop, vital signs are less likely to be recognised (McGregor et al, 1998).

3.6 The decision to call emergency services for help during an overdose may be influenced by past experience. According to Tobin and colleagues (2005), past witnesses of a fatal overdose were almost twice as likely to call emergency services compared to those who had been present at a non-fatal overdose. They suggest that witnessing a fatality may "sensitize drug users to the seriousness of overdose".

3.7 However, the study paints a more complex picture. The authors suggest that drug users who have survived overdose and go on to witness it may be a) less aware of the life-threatening nature of the situation and b) less likely to call an ambulance. They may also feel more confident and competent in managing the overdose situation.

3.8 Commenting on overdose casualties that received emergency medical help, the authors hypothesise that their experiences and perceptions of this help may reduce the future likelihood of calling an ambulance (Tobin et al, 2005).

3.9 A New York study by Tracey and colleagues (2005) found that witnesses of overdose events within public areas were more likely to summon medical help compared to overdoses occurring within private locations. Furthermore, those taken to hospital following a recent overdose were more likely to call for help than those who had not. Fears about medical care and police involvement (which are common barriers to seeking help) may be less acute among those who experienced an overdose and subsequent hospitalisation.

3.10 The presence of bystanders may decrease the likelihood of calling an ambulance. To reduce the likelihood of a "diffusion of responsibility", Tobin and colleagues (2005) suggest that drug users should be trained to direct someone present to be responsible for calling an ambulance while others attempt resuscitation. Noting that drug users tend not to telephone an ambulance as a first response (Fitzgerald, 2000), it has been suggested that two components require consideration - diagnosis and intervention (Best et al, 2002).

3.11 Many initial witness responses, such as slapping and shaking the casualty, may be attempts to assess the severity of the problem and their capacity to manage the situation before considering external help. Best and colleagues (2002) found that remaining with the casualty may help prevent choking or provide a level of sensory stimulation that prevents them falling too far into an overdose state. Thus the continued presence of witnesses attempting a range of resuscitation methods may play a critical role in the prevention of many fatalities even if some of these techniques are individually ineffective.

3.12 Witnesses who attempted CPR prior to ambulance arrival improved hospitalisation rates compared to cases where it was not administered (Dietze et al, 2002). Moreover, CPR administration was associated with a statistically significant improvement in clinical outcomes in cases of non-fatal heroin overdose; for example, complications due to prolonged depression of respiratory function and conscious state. The authors suggest that benefits might include a reduction in the incidence and severity of cases of hypoxic brain injury (Dietze et al, 2002).

3.13 Pollini and colleagues (2006) have noted that intervention was also more likely when witnesses had received information on how to prevent/revive a casualty compared to those who had received no information. This reinforces the views that providing relevant information may be an effective strategy to help prevent or reduce further harm such as related morbidity and deaths (Dietze et al, 2002; Bennett et al, 1999; Best et al, 2002; Tobin et al, 2005; Zador et al, 1996; Wright et al, 2005; Hall, 1998). Additionally, offering CPR training and other interventions (such as naloxone or emphasising witnesses to remain with the casualty until medical help arrives) should be offered to those likely to be present, such as drug using peers, family and friends.

Ambulance

3.14 An Austrian study conducted in Vienna used ambulance service data to observe illicit opiate use. The study reviewed the records of ambulance crews called out to emergencies where a diagnosis of heroin or opiate overdose was recorded over a 14 month period (Seidler et al, 2000). With 707 people involved in 1087 non-fatal overdoses, an important sub-group was identified - more than half (52%) of all the 1087 emergencies were attributed to 189 drug users.

3.15 This Viennese ambulance data was used by local drug agencies to target services by mapping high risk areas and identifying those repeatedly overdosing. It also lead to the identification of high-risk groups (in this case young people) not previously identified by services with subsequent help being offered, which included drug counselling.

3.16 Seidler and colleagues argue that this ambulance data allows for rapid discussions and focussed attention. They also state that undertaking a local evaluation of emergency service responses, which includes operational structures, could result in a new and useful source of information on drug use and drug-related deaths.

3.17 In Melbourne, Australia, a similar approach was undertaken which involved establishing a database of non-fatal heroin overdoses attended to by ambulance personnel (Dietze et al, 2000). The aim of this ambulance database was to provide interested stakeholders with reliable, quality and up-to-date data on heroin-related harm. Similar to the work in Vienna, the Melbourne data was used to map high-risk areas and identify overdose clusters within a number of areas. Although police attendance is often cited as a key barrier to people not contacting ambulance services, Dietze and colleagues (2000) noted low police attendance at drug overdose scenes (12%). They also stated that strong links developed between researchers and the ambulance service, may serve as the basis for important future research regarding heroin overdose.

3.18 In the UK, extensive work carried out on behalf of the Joseph Rowntree Foundation ( JRF) emphasised the need for standardised ambulance call-out statistics collated at a national level. The JRF Independent Working Group ( IWG), which examined international work on the role of Drug Consumption Rooms, stated that the lack of data on fatal and non-fatal drug overdoses was a "significant weakness" in the evidence base (Joseph Rowntree Foundation, 2006).

Police

3.19 Several authors have reported that a main barrier to calling for help is the "fear" of police involvement (Pollini et al, 2006; Tobin et al, 2005; Tracy et al, 2005; Bennett et al, 1999). It has been suggested that this barrier requires research attention to provide a better understanding of drug users' fear of arrest and how barriers can be reduced (Tobin et al, 2005). It has also been argued that liaison between police and ambulance services, aimed at supporting the safe calling of ambulances - should form a component of any planned intervention (McGregor et al, 1998).

3.20 At a structural level, reducing police attendance at the scene of an overdose and decreasing the risk of arrest might increase willingness to call emergency services (Pollini et al, 2006; Bennett et al, 1999).

3.21 Although not formally evaluated, some UK police force areas have been involved in developing protocols regarding police attendance at overdose incidents. For instance, an agreement was reached between Nottinghamshire Police, the East Midlands Ambulance Service and the local DAATs to ensure that police officers do not routinely attend ambulance call-outs to drug overdoses unless a death has already occurred; there are child protection concerns; and/or the address is identified as one where there could be a threat of violence. Similar protocols have been established in other parts of England, such as Kirklees, Leicestershire and Avon & Somerset.

Accident and Emergency

3.22 A recent unpublished Edinburgh study, which analysed 90 drug-related deaths over a two-year period in the Lothian region, investigated the association between drug-related deaths and past contact with the Royal Infirmary of Edinburgh (Thanacoody et al, 2007). Just over half of the confirmed deaths had previous hospital contact within five years of their death. More than one third had contact within 12 months of their death.

3.23 A similar study examined methadone-related deaths in the Lothians between 1997 and 1999 (Fiddler et al, 2001). It noted that 60% had attended accident and emergency departments for deliberate self-harm or accidental overdose. Commenting on this finding, the authors described these periods of hospitalisation as providing a "unique" opportunity for appropriate interventions to be targeted at these high-risk patients.

3.24 Other European studies have also identified missed opportunities for intervening within medical settings (Pollini et al, 2005; Cook et al, 1998). The studies noted that the number of patients receiving treatment information from emergency departments or hospital staff was low, as were the numbers referred on to drug treatment. Importantly, Pollini's study found that hospital staff and crisis counsellors appeared particularly influential in linking injecting drug users ( IDUs) with drug treatment.

3.25 Clearly, there is a need for medical care providers to capitalise on contact with drug users following an overdose event and provide information on overdose prevention strategies and referral to drug treatment programmes. Thanacoody and colleagues (2007) point to liaison between emergency departments, clinical toxicology services and community drug-based addiction services to help increase the number of drug users engaging with community treatment services. Other broader policy suggestions include routine screening for health-damaging behaviours and implementation of health promotion strategies within general hospitals (Canning et al, 1999).

3.26 In Scotland, innovative practice is being developed to meet these challenges. For example, Monklands Hospital, in NHS Lanarkshire, have located substance use specialist nurses within the accident and emergency department to progress referral for drug overdose casualties, and provide advice and information to family members or significant others accompanying the casualty.

Communicating public health alerts

3.27 Appropriate dissemination of health messages may be an important vehicle for reducing drug-related deaths. Therefore, it is important that identification of drug users' information networks is explored.

3.28 An investigation has explored the communication channels through which drug users receive information on "bad dope" (Freeman and French, 1995). In this instance, it was Fentanyl (an extremely potent opioid analgesic) which contributed to a small number of fatalities and an increase in hospital admissions in New Jersey. There were geographical differences in sources of information across three city areas in New Jersey with notable gender differences. Male drug users were most likely to have received their information from the TV while females were more likely to have heard from friends. Other information sources included radio, newspapers, other drug users, relatives and to a lesser extent police sources. The authors cautiously advise that public health alerts could have the paradoxical effect of increasing some users' interest in obtaining a particular drug.

3.29 Elsewhere, media coverage that refers to street heroin locations has been linked to the increased use (Fitzgerald, 2000). When asked about trusted sources on providing good information about "bad dope", friends and other addicts were considered most reliable with no one regarding TV, radio or the police as reliable sources.

3.30 These papers suggest that health officials need to understand how public health messages are perceived and processed by drug users and should include further exploration of those sources considered trustworthy.

Section 2 - Mainstream & Emerging Interventions

Drug Treatment and Methadone

3.31 A small Scottish study of 33 drug overdose casualties attending 6 accident and emergency departments in 2 Scottish cities may provide some drug treatment, policy and practice insights (Neale, 2000). The researchers identified 4 overdose situations related to methadone and methadone treatment: 1) Topping up on a legitimate methadone prescription 2) Using someone else's methadone prescription 3) Preferring illegal drug use in favour of prescribed methadone 4) Unable to access a methadone prescription.

3.32 Methadone diversion was viewed as an important factor contributing to non-fatal overdose which was common among those already prescribed methadone. The author considered tighter supervision of methadone consumption in pharmacies and drug clinics as a way of reducing illicit diversion.

3.33 Conversely, Neale (2000) noted that methadone-related overdoses occurred among those unable to obtain substitute medication, despite a number often having had previous methadone prescriptions. Careful monitoring and evaluation of substitute prescribing should include the opinions and concerns of the drug users by actively involving them in their treatment decisions wherever possible (Neale, 2000).

3.34 Some overdose casualties had not always taken their prescribed medication which may have prompted reduced drug tolerance, withdrawals and an increased susceptibility to overdose. Those casualties that consumed methadone prior to overdose cited a range of explanations - from unintentionally taking too many drugs, unexpected heroin purity to a lower tolerance or ingesting unknown tablets. The author suggests that despite drug users understanding the risks, more information is required as to why "self-destructive" behaviours persist. With a significant number requesting additional support, those leaving hospital should receive follow-up support and/or additional help to avoid future drug overdose (Neale, 2000).

3.35 Echoing some of these Scottish findings, an NTORS study found that clients who had overdosed in the three months prior to treatment were more frequent users of illicit methadone (Stewart et al, 2002) and others found that most people involved within a structured methadone maintenance programme reported continued illicit drug use (Cullen et al, 2000). Therefore, the risks of consuming both prescription and illicit drugs need to be addressed among users and service providers.

3.36 Stewart and colleagues (2002) found that the continued incidence of overdose among some clients at one year follow-up was a cause for concern. In accordance with these findings, a study in Dorset linked overdose to very high levels of drug intake with users experiencing difficulty in controlling their drug intake (Bennett et al, 1999). Drug injectors at greater risk, and in contact with mainstream drug services and prison, are in an ideal position to be offered overdose prevention work (Bennett et al, 1999).

3.37 The relationship between poor treatment response and non-fatal overdose suggests there is a strong case for incorporating non-fatal overdose into routine measurements of treatment in order to target interventions at individuals most at risk. Fischer and colleagues (2004) suggest further research is required to disentangle the "complex dynamics of the potential anti-therapeutic effects of treatment", taking into account the fact that many treatment episodes for drug users are suddenly or prematurely terminated, with no opportunity for transition measures.

3.38 Addressing some of these challenges, a National Treatment Agency ( NTA) briefing paper examined the evidence on methadone dose and maintenance treatment. A key research message was the consistent finding of greater benefit being accrued from offering most individuals on methadone maintenance a daily dose between 60mg and 120mg. Yet the paper noted that British methadone treatment doses are on average less than 50mg daily with only one in four service users receiving over 60mg ( NTA, 2004).

3.39 The NTA briefing paper also found that higher doses were consistently shown to encourage treatment retention and reduce illicit drug use in methadone maintenance regimes. Conversely, lower dose levels may undermine the provision of optimal services and compromise the therapeutic relationship between service user and key worker. The briefing paper also noted that responsive and flexible individualised dosing can help foster the therapeutic relationship, and lead to improved outcomes and reductions in illicit drug use ( NTA, 2004).

General Practitioners

3.40 Despite many General Practitioners ( GPs) playing an active role in the management of drug problems, including prescribing substitute drugs such as methadone, there are limited studies looking at the role of the GP in the management and prevention of drug-related overdose.

3.41 An Irish study involving a small sample of heroin users in a Dublin GP practice revealed high levels of activity associated with overdose and poor preventive measures (Cullen et al, 2000). Although the sample had significant personal experience of overdosing or knowing people who had died, there was poor knowledge of preventative measures and how to manage an overdose. The majority of this sample was involved in a structured methadone maintenance programme, but reported continued illicit drug use and ongoing exposure to witnessed overdoses.

3.42 This Dublin study found that GPs recognised the importance of being involved in blood borne virus ( BBV) interventions with drug users. However, their role in responding to overdose activity was not well recognised. Cullen and colleagues recommend that overdose prevention and management should become a priority for GPs caring for opiate-dependent patients.

3.43 An Australian study looked at prescription drug-seeking behaviours among young people who died of heroin-related overdose (Martyres et al, 2004). Key study findings included high levels of poly-drug use and prescription drug use among the heroin deaths and circumstantial evidence of increasing use of multiple doctors and excessive increases in psychoactive drug prescriptions.

3.44 Increased GP attendance may be an "indicator" of overdose risk but also an opportunity to intervene and advise injecting drug users about treatment options (Martyres et al, 2004). However some GPs may be reluctant to become involved in identifying and managing drug users and when faced with persistent and threatening patients, the temptation to prescribe on request may be an easier option. Martyres and colleagues suggest that there is a need for a longitudinal study of heroin users, in relation to fatal and non-fatal overdose, to assess if increased "doctor shopping" is a predictor of overdose risk.

The impact of services engaging and retaining drug users

3.45 A study by Digiusto et al (2004) noted that all deaths and most overdoses occurred after leaving treatment. Other authors have also pointed out that those engaged in treatment were at lower risk of death (Fugelstad et al, 2007; Darke et al, 2005; Bartu et al, 2004). A ten-year longitudinal mortality study found no significant differences between two treatment types (methadone versus buprenorphine maintenance treatment) but concluded that increased exposure to maintenance treatment decreases the risk of death (Gibson et al, 2008).

3.46 In an Australian study, the number of heroin users who overdosed declined by half following enrolment in treatment; with the risk further reduced the longer people stayed in treatment (Darke et al, 2005). The study identified that a greater number of separate treatment episodes lead to an increase in overdose risk, leading the authors to highlight the importance of treatment stability, longer spells in services and less treatment episodes to improve outcomes (Darke et al, 2007).

3.47 A Swedish study examined opiate users who had been in contact with a methadone treatment programme, from 1988 to 2000, which included those discharged from treatment and those not accepted into treatment (Fugelstad et al, 2007). It found the lowest mortality rates among those within the methadone treatment programme and the highest rates among those who had left the service or were discharged from it.

3.48 This Swedish study also highlighted that different countries choose different strategic approaches towards methadone programmes - from "low threshold" programmes that prioritise availability and try to keep people in treatment for as long as possible to "high threshold" approaches that prioritise security. Although the Swedish methadone programme protected people from fatal heroin overdose or methadone intoxication, it was not attractive or easily accessible to many and there was a high mortality rate among those discharged from the programme. The authors concluded that different treatment polices and rules of inclusion lead to different mortality patterns with strict rules increasing the risk of discharge resulting in a high mortality rate (Fugelstad et al, 2007).

3.49 Examining mortality rates after one year among people in a methadone treatment programme, Zanis and Woody (1998) found that discharged patients were 8 times more likely to be dead compared to those still in treatment with the main cause being drug-related overdose. Although it was not possible to know if those discharged would still be alive if they had remained in treatment, the authors noted that the significant differences in mortality would imply that it may have produced a more favourable outcome. They suggest the need for more tolerant programmes to increase retention among less compliant active drug users; restrict the number discharged due to on-going drug use; and, explore alternative ways to reduce drug use. Other suggestions include follow-up and assessment of those discharged to provide opportunities to re-enter treatment or enrol in other programmes (Zanis and Woody, 1998).

3.50 An Italian study examining unintentional illicit drug overdose between 1984 and 2000, found that withdrawal from drug treatment was an important precursor to fatal overdose - most deaths occurred among those out of treatment for more than two weeks (Preti et al, 2002). The authors concluded that the greater availability of drug treatment services in Italy may have been partly responsible for the decrease in the risk of death by overdose among injecting users during the study period.

3.51 Another Italian study drew attention to the importance of retention for long-term and maintenance clients as a means of preventing overdose (Davoli et al, 2007). The authors found that the risk of overdose within the first 30 days after stopping/completing treatment was 3 times higher compared to 31 days or more after treatment. They also identified an increased mortality risk among those that finished methadone detoxification compared to those who had ceased or dropped out of it. This was attributed to greater reductions in tolerance among those finishing detoxification thus increasing overdose risk following relapse. The authors emphasised the importance of adequate follow-up among abstinence-based treatment providers and educating drug users about the risks of post-treatment relapse and overdose (Davoli et al, 2007).

3.52 Exploring mortality among opiate and amphetamine users in Perth (Western Australia), Bartu and colleagues (2004) found that participants engaged in treatment are at lower risk of death regardless of the treatment received. Those opiate users that withdrew from treatment were more than 8 times at risk of drug-related death, 6 months after treatment. Those who withdrew from treatment against advice were also at higher risk. The authors emphasised the need for clinicians to stress that those withdrawing from treatment can return at any time in order to minimise the risk of death should they relapse.

3.53 A study in London into the characteristics and types of overdose deaths endorses these other European findings. Hickman and colleagues (2007) suggest that increased methadone prescribing was one explanation for the overall decline in drug-related deaths in England and Wales between 2000 and 2003. Therefore, they argue that increasing the availability of treatment among heroin users both in the community and in prison is vital to reducing drug-related mortality rates.

3.54 Other treatments, such as heroin-assisted treatment may also reduce mortality rates, according to Rehm and colleagues (2005). The Swiss study of mortality rates among those involved in heroin-assisted treatment between 1994 and 2000 found that the rates among those in this treatment were lower than that of other users both in and out of treatment.

3.55 Although treatment retention is an important protective factor, Darke and colleagues (2005) stress the need to consider other important risk factors such as polydrug use and recent overdose.

Emerging interventions - Take-Home-Naloxone and Safer Injecting Rooms

Take-Home-Naloxone

3.56 With most drug overdose deaths occurring in the company of others, there are opportunities to intervene using naloxone (Strang et al, 1999; Lenton and Hargreaves, 2000; Baca and Grant, 2005). Naloxone, an antagonist drug used to reverse opioid overdose, has been identified as the single most important resuscitative action during heroin overdose (Baca and Grant, 2005). Reinforcing this view, Strang and colleagues (1999) argue that at least two-thirds of the 69 overdose fatalities identified in their study could have been prevented by immediate administration of take-home-naloxone ( THN). At risk groups that could benefit from THN are detoxified opiate users discharged back into the community, those in the first few weeks of methadone substitution therapy, and opiate users being released from prison (Strang et al, 1999).

3.57 A national naloxone project involving emergency services, clinicians, and clients was carried out by the National Treatment Agency ( NTA) in England (Strang et al, 2007). After being trained in overdose management, 239 clients received a THN supply.

3.58 NTA follow-up of 186 (78%) THN clients revealed that 18 overdoses were witnessed and 10 naloxone administrations were carried out with no adverse consequences and full success in overdose reversal. The study uncovered high rates of personal/witnessed overdose among opiate users attending treatment services but also high levels of support for expanding the provision of THN to prevent fatalities. Although there was scope for improving awareness of overdose prevention and naloxone administration, the study noted differences in the extent to which services were willing to commit time and resources to this THN initiative. However, clients who had used naloxone expressed a commitment to the project and suggested the biggest challenge was continuing to raise overdose awareness and provide training (Strang et al, 2007).

Take-Home-Naloxone concerns

3.59 It has been suggested that THN may encourage a small minority to increase their drug use, use in a more risky way (Strang et al, 1999; Lenton and Hargreaves, 2000) or take more heroin to lessen naloxone-induced drug withdrawals thus potentially falling back into a state of overdose (Worthington et al, 2006; Seal et al, 2003). This latter point was explored in a study examining injecting drug users' ( IDUs) attitudes towards being prescribed THN - 46% stated that they might not be able to dissuade the casualty from using more heroin following THN administration (Seal et al, 2003).

3.60 With the drug having a short-acting duration (30 to 90 minutes) more than one dose may be required when long-acting drugs, such as methadone, have been used (Baca and Grant, 2005; Lenton and Hargreaves, 2000). Offering THN to opiate users could have significant health implications as injecting naloxone could potentially increase the transmission of infectious diseases (Baca and Grant, 2005). There is also the probability that drug users and their peers offered THN may be less likely to call an ambulance resulting in fewer non-fatal overdose casualties being medically reviewed with associated morbidity remaining undetected and untreated (Lenton and Hargreaves, 2000).

Addressing Take-Home-Naloxone concerns

3.61 It has been argued that THN is a safe intervention and fears regarding its use are not well-founded (Baca and Grant, 2005). Lenton and Hargreaves (2000) emphasise that no significant problems have arisen following hundreds of administrations in both the UK and Australia and note that similar concerns that were raised about needle exchanges have proven unfounded. They also point to follow-up THN research in Berlin - involving a programme set up in 1999 - which did not identify any cases of risky drug consumption. Furthermore, the abuse potential is considered negligible as naloxone has no reinforcing properties and rapidly provokes unpleasant withdrawal symptoms thus reducing the likelihood of abuse (McGregor et al, 1998).

3.62 The lack of reinforcing properties were evident in a New York study which suggested that drug users were unlikely to engage in riskier drug-taking activity (Worthington et al 2006). Those with experience of administering THN described the incident as challenging, stressful and emotionally upsetting with some put off by the potential for "dopesickness" (or opiate withdrawal) after THN administration. However, there were no reports of study participants refusing to seek medical help after THN administration. The authors conclude that widespread THN availability would not weaken the important message of contacting the emergency services following overdose (Worthington et al, 2006).

3.63 Addressing the potential transmission of infectious diseases, Baca and Grant (2005) suggest that medical staff could combine naloxone distribution with syringe exchanges and user education regarding blood borne virus ( BBV) transmission with THN programmes offering prevention and treatment opportunities to high-risk drug users. Other concerns could be reduced by looking at alternative methods of administering naloxone

3.64 Recently, Kerr and colleagues (2008) examined the use of intranasal naloxone for the treatment of heroin overdose. They found it to be a safe and effective option, which could be useful for administration within communities as it would reduce the risk of needle stick injuries for care-givers and reduce discomfort for those receiving it. Despite these advantages, they emphasise that there is still a lack of evidence to support its use as a first-line intervention by paramedics for the treatment of heroin overdose and call for further research to verify its effectiveness, safety and value.

3.65 With improving witness response a major challenge, the study by Lenton and Hargreaves (2000) found that, in practice, witnesses only called an ambulance in about one in 10 overdose incidents with no reported intervention taking place in nearly 8 out of 10 deaths. Addressing some of these concerns, Worthington and colleagues (2006) suggest that THN may prevent significant others reverting to potentially harmful and less effective resuscitation methods, but this will require increased education and resources.

3.66 In New York, drug users undertaking THN programmes reported gaining confidence in administering the drug through practice and follow-up training. Commenting on this work, Worthington and colleagues (2006) believe programmes need to arrange multiple visits with enrolled participants to practice role play in administering the drug, offering them support and addressing their fears. This may lessen the detrimental effect of panic and intoxication on successful THN administration.

3.67 Putting forward the view that peers are more likely to know what drugs the person has taken, monitor their initial response to THN and administer a subsequent dose if necessary, Lenton and Hargreaves (2000) suggest a range of measures: from encouraging peers to seek medical help to providing them with controlled amounts of methadone or buphrenorphine to ensure the casualty experiences some relief from drug withdrawal. Others suggest the need to emphasise strategies within overdose prevention programmes that ensure effective response to potential THN risks (Seal et al, 2003). Commenting on some of the insights gained from New York City's THN programmes, Piper and colleagues (2007) conclude that programme experiences and data shows that these initiatives are a feasible option in effectively training drug users to respond effectively to overdose by administering THN. The authors emphasise the need for flexibility and simplicity in the development, implementation and evaluation of these types of programmes, adapting them to suit the needs and experiences of participants. Moreover, it was also considered important to incorporate user feedback in the planning and delivery stages (Piper et al, 2007).

3.68 In summary, there is a consensus among the reviewed papers that there is a potential to prevent many opiate overdose deaths using THN. The possible benefits of THN are considered sufficient to justify the need for carefully monitored pilot schemes that are linked into extensive educational programmes and training (Strang et al, 1999). ²

Safer Injecting Rooms

3.69 Advocates of safer injecting rooms ( SIRs) claim that these facilities can help reduce harms associated with IDU, such as heroin overdose levels (fatal and non-fatal), BBV transmission and the impact of street-based injecting. In a study carried out in Melbourne, prior to the establishment of a SIR, participants were aware of SIRs and their main components (Craig, 1999). A number of participants expressed concerns about injecting in public spaces and the risks of heroin overdose - they stated that SIRs had an important role to play here. Those participants who reported being most willing to use SIRs were male, had experienced more non-fatal heroin overdoses and used heroin more frequently in the 6 months prior to interview, compared to those not willing to use SIRs. The former group would be an important target group for harm-minimisation strategies, such as SIRs.

3.70 This Melbourne study also reported that a significant number of IDUs were not willing to use SIRs. Reasons cited included a preference to use in a private setting. Yet, the report authors point to evidence that shows that most heroin overdoses (fatal and non-fatal) occur within a private setting. If SIRs were established, it may be that the risk of overdose mortality and morbidity would likely continue among a significant number of this group who prefer injecting in private (Craig, 1999).

3.71 A literature review by Hunt (2006) for the IWG on Drug Consumption Rooms suggested SIRs can contribute to a reduction in drug-related deaths, although the significance of their effect depends on variables, such as the extent to which they reach their target population and the number of deaths occurring outside that target population - for example, those who use in private and among more socially integrated users. Nevertheless, there is no evidence that the use of SIRs contributes to increased risk of morbidity or mortality. Hunt reported that no fatal overdoses has occurred within a SIR despite there being "millions" of supervised drug consumptions and thousands of treated emergencies, thus showing evidence that SIRs provide a high level of safety from overdose among the people who use them.

3.72 An illustrative example of SIRs' beneficial effects is the EVA project in Barcelona (Anoro et al, 2003). Records from the EVA project (from January 2001 to March 2003) showed that staff assisted 377 cases, 52% of which involved respiratory arrest, with no overdose deaths occurring during EVA opening hours or within the larger community. Eight out of 10 overdose interventions were carried out by EVA nursing staff with less than one out of 10 cases requiring an ambulance call out.

3.73 According to the authors, the availability of naloxone for staff and clients significantly helped to reduce overdose mortality rates, with staff operating within strict CPR/naloxone protocols. The EVA project also facilitated recruitment and training of active drug users in basic CPR which included providing them with THN. This take-home initiative was estimated to have reduced overdose mortality by one third in the Can Tunis area of Barcelona between 2000 and 2001.

Section 3 - Early Individual & Social Indicators

Tolerance

3.74 Numerous national and international studies have identified that reduced tolerance to opioids is a major risk factor in heroin-related overdose deaths. A study in Sheffield, examining the role of concomitant drugs and risk factors in accidental fatalities between 1997 and 2000, found that one in 5 deaths were after a period of abstinence from regular use, suggesting that decreased tolerance is a key factor (Oliver and Keen, 2003). The most frequently reported reasons for abstinence were imprisonment and hospital admission. The authors highlighted that research into fatal overdose following release from prison has been conducted on several occasions but that there is a need for further research to fully assess the risk of fatal overdose faced by opiate misusers discharged from hospital.

3.75 Examining drug related mortality for male ex-prisoners between the ages of 15 and 35 years old, Bird and Hutchinson (2003) found that it was 7 times higher in the 2 weeks after release than at other times of liberty. They estimated one drug-related death in the 2 weeks after release per 200 adult male injectors incarcerated for 14 days or more.

3.76 A study conducted by Jones and colleagues (2002) of drug users in Glasgow who had died of fatal overdose highlighted that, although the Scottish Prison Service had tried to tackle this problem by providing pre-release information about overdose risks and arranging for continued support from community drug services, drug users remanded in custody or released at short notice were likely to miss out on this support, suggesting the need to also target those facing this situation.

3.77 The results of a study by Thiblin and colleagues (2004) into heroin-related deaths in Stockholm between 1997 and 2000 as a result of intranasal administration (snorting) and pulmonary inhalation (smoking) also found that reduced tolerance is a major risk factor. The study revealed that low levels of tolerance are of particular significance in cases of heroin-related death involving administration routes other than injection. The study highlighted that, although these forms of administration are generally documented to be less risky, it is important to be aware that they lead to highly variable blood morphine concentrations and, thus, do not protect against lethal intoxication. The majority of individuals examined in this study were trying to reduce their level of drug use and using heroin less frequently thus indicating that low tolerance may have been an important factor in the fatal outcome for these non-injectors. This study supports other evidence that when tolerance has been lowered, rather than protecting against fatalities, the sporadic use of heroin is a major risk factor regardless of the chosen method of use (Thiblin et al, 2004).

3.78 Several studies have begun to identify the risks involved in methadone maintenance treatment. Rugelstad and colleagues (2006) state that methadone is not only a "life-saving" drug but can also be a "fatal" drug. Wolff and colleagues (2002) found that overdosing with the drug has become more common and, although little is known about the circumstances surrounding methadone deaths, some of the people at highest risk are those whose usual tolerance has been reduced. Others have pointed out that the risk of overdose is generally higher during periods of induction and transition, such as when drug users (re)enter or discontinue treatment (Bell and Zador, 2000; Buster et al, 2002). For instance, fatal outcomes are often the result of prescription doses that exceed the user's tolerance level (Bell and Zador, 2000) with higher overdose fatalities occurring during the first 2 weeks of treatment (Buster et al, 2002). There is a need for adequate assessment and review of tolerance prior to treatment among new and returning patients seeking help, especially recently liberated prisoners (Bell and Zador, 2000; Buster et al, 2002).

3.79 Wolff and colleagues (2002) have pointed out that drug tolerance develops at different rates and is often moderately slow with methadone. Therefore, problems may arise if the person's dose is increased too quickly, or if the initial dose is too high. The need for tolerance testing (Wolff et al, 2002) is reinforced by the view that newly inducted methadone clients should be monitored closely during the initial days of treatment (Bell and Zador, 2000).

The impact of injecting drug use, benzodiazepines, alcohol and cocaine

Injecting drug use and benzodiazepines

3.80 The major NTORS and DORIS studies note that injecting drugs and benzodiazepine use are major factors in drug-related overdose (Stewart et al, 2002; Neale et al, 2005). Although one study found that benzodiazepine use increased the risk of overdose 28 fold (Dietze et al, 2005), of more concern the NTORS study indicated that 9 out of 10 drug injectors entering treatment reported past overdose. This group of injectors were also 10 times more likely to have overdosed than non-injectors entering treatment and were still the biggest at-risk group, at one year follow-up.

3.81 The NTORS one-year follow-up found that reductions in overdose were closely linked to large reductions in rates of injecting behaviour. Those reporting problems associated with injecting, such as abscesses or poor injecting practices, were significantly more likely to report an overdose. The authors suggest that interventions directed at these health problems may provide a useful opportunity to include information and counselling designed to reduce overdose risk (Stewart et al, 2002).

Alcohol consumption

3.82 A number of studies have reported alcohol consumption as being an overdose factor (Zador et al, 1996; Gossop et al, 2002; McGregor et al, 1998). The NTORS study found that clients drinking large quantities of alcohol were at greater risk of overdose. Failure to address their alcohol problems meant a continued risk of overdose despite improvements in levels of drug use after treatment (Stewart et al, 2002). A study of street-recruited heroin injectors in San Francisco Bay identified important independent risk factors which included being younger, frequently arrested, participation in methadone detoxification but also moderate to heavy daily alcohol consumption (Seal et al, 2001).

3.83 Targeting interventions at clients identified as daily alcohol users and those who are frequently arrested may help reduce the frequency of non-fatal and fatal overdoses among this particular group of drug users.

Cocaine

3.84 There has been relatively little research conducted into patterns of cocaine overdose and its contribution to overdose mortality and morbidity. Among the few papers looked at for this review, cocaine overdose was more common among injecting cocaine users ( ICU) (Bernstein et al, 2007; Kaye et al, 2004; Pottieger et al, 1992). Females were also more likely to report a cocaine overdose and, as with opiate overdoses, long-term users were more likely to experience a cocaine overdose than younger users reflecting perhaps prolonged risk exposure or the cumulative effects of cocaine, which increases the risk of a toxic reaction over time (Bernstein et al, 2007; Kaye et al, 2004).

3.85 Injecting cocaine users ( ICU) are reportedly more likely to have witnessed a cocaine overdose (Kaye et al, 2004) and although interventions to reduce opiate overdoses have gained importance and wide support, the findings from Kaye and colleagues suggest that drug users possess a poor knowledge of cocaine overdose and appropriate interventions. Moreover, given the paucity of UK research into cocaine overdose, relatively little is known about the prevalence of and risk factors associated with cocaine overdose. If these findings hold true in Scotland, it is vitally important that measures are put in place to increase cocaine users' knowledge about the risks of cocaine overdose and appropriate responses to them.

Health Morbidity

3.86 Despite the fact that there appears to be extensive health morbidity associated with non-fatal overdose, it remains a relatively unexplored area. Warner-Smith and colleagues (2002) categorised overdose sequelae as follows: pulmonary effects, such as oedema, often resulting in pneumonia and occurring as a common complication of aspiration; cardiac effects, such as arrhythmia, acute cardiomyopathy and haemoglobinaemia; muscular effects, such as rhabdomyolysis; and neurological effects, such as cognitive impairment, resulting in impaired judgement regarding polydrug use, doses and tolerance.

3.87 In one of the first organized attempts to explore the extent of overdose-related harm among heroin users in the Sydney region in Australia, Warner-Smith and colleagues (2002) found that over three-quarters of overdose casualties had experienced at least one morbidity symptom. The most commonly direct symptoms were peripheral neuropathy, due to prolonged pressure on limbs while unconscious and pulmonary complications; commonly reported indirect injuries were from falling and burns. It is noteworthy that the study found that morbidity is a universal problem for all heroin users - no meaningful differences were identified between treatment and non-treatment groups.

3.88 With older long-term dependent users most at risk of fatal opioid overdose, Darke and colleagues (2006) carried out an analysis of coronial cases to explore the relationship between age and overdose. Warner-Smith and colleagues (2001) also identified age as a major problem, stating that the morbidity burden is directly related to the number of overdose episodes experienced and is therefore more likely to be greater among older, more experienced and more dependent users. Moreover, Bartu and colleagues (2004) noted that because the age at which people begin using drugs is falling, the length of time that people have used drugs may be a stronger indicator of overdose rather than chronological age.

3.89 The 2006 study by Darke and colleagues found that multiple systemic diseases, in particular hepatic and cardiac disease, are prominent among older cases - being found in nearly half of the oldest age group of the sample. The study identified high levels of hepatitis and cirrhosis which may have a serious effect on the drug user's ability to metabolise opioids and poor cardiac health which may increase the risk of hypoxia-induced cardiac arrest and arrhythmia. Webb and colleagues (2003) also identified an increasing risk of death among older users, and methadone users in particular, arguing that they are at high risk of contracting acute infections leading to septicaemia or endocarditis or contracting chronic infections such as HIV, Hepatitis B and Hepatitis C.

3.90 Several authors believe that the progressive disease burden acquired as drug users grow older means that they are more susceptible to overdose and that there is a clear case for regular medical examinations and liver function tests (Darke et al, 2006; Warner-Smith et al, 2001; Warner-Smith et al, 2002).

Recent Life Problems/Psychological Factors

3.91 Using data from the DORIS study, Neale and Robertson (2005) investigated the role of recent life problems in non-fatal overdose among heroin users entering various drug treatment settings. Results from the analyses suggested individuals who had overdosed recently (90 days prior to interview) were significantly more likely than the other participants to have recently experienced bereavement of someone close to them, a relationship breakdown and to have had accommodation problems. However, the authors acknowledge it was not possible to say whether the variables associated with recent overdosing occurred before or after the overdose incidents, thus they cannot be sure the life problems were instrumental in causing overdose. Nevertheless, they suggest that service providers should seek to identify and address drug users' problems as part of a broad strategy of overdose prevention. Assessment tools could cover a range of personal and social issues including a measure of suicidal intent. Similarly, clients should be encouraged to talk openly about any problems affecting them. Furthermore, the findings support the need for a coordinated approach among treatment agencies and other health, social care and criminal justice professionals, as well as specialist counsellors relating to bereavement and relationships. The findings also support the need for a coordinated approach particularly among community drug services and residential detoxification units where recent life problems were associated most strongly with recent overdose.

3.92 Similarly the NTORS study found that non-fatal overdose was also associated with more self-reported psychological health problems at treatment intake and at one year follow up (Stewart et al, 2002). Levels of anxiety were predictive of non-fatal overdose both before treatment and at one-year follow-up. A greater proportion of clients reporting a non-fatal overdose before treatment also reported suicidal thoughts with suicidal ideation a predictor of overdose at one year.

3.93 A recent case-control study which involved carrying out a "psychological autopsies" found that in cases where fatal non-deliberate overdose occurred, the person was more likely to have a history of mental health problems, a current psychiatric diagnosis and to have been prescribed psychotropic medicines (Oliver et al, 2007).

Antidepressants

3.94 Oyefeso and colleagues (2000) highlighted that access to anti-depressants, through genuine prescriptions, is a prominent risk factor for fatal anti-depressant overdose ( FAO). Despite certain limitations, the results of Oyefeso's study into the predictors, extent and pattern of FAOs in England and Wales showed that this risk is particularly prominent among female drug users. It highlighted that this fact, together with the high risk of suicidal intent among female addicts and non-addicts and the identification of a subgroup of female drug users with mood disorders, suggests a need to obtain and carefully evaluate drug-abuse history of women with an affective disorder to reduce the risk of antidepressant misuse.

3.95 Cheeta and colleagues (2004) identified that deaths in which antidepressants are implicated in combination with other drugs are more likely to involve drug users and that the drugs most commonly implicated are alcohol and opiate-based drugs. These findings indicate that all patients prescribed antidepressants should be routinely screened for a history of problem drug use, and that GPs and psychiatrists should display caution when prescribing for these individuals. This is of particular relevance taking into account the fact that the National Investigation into Drug Related Deaths in Scotland published by the Scottish Government in 2003 identified that 36% of the cohort were prescribed anti-depressant medication in the 6 months prior to death with GPs being the main prescriber. Furthermore, Oyefeso and colleagues (2000) identified that one risk factor associated with potential fatalities involving antidepressants is that patients may obtain different antidepressants from different prescribers; a practice that is commonly associated with drug users and known as "doctor shopping" (Martyres et al, 2004).

3.96 Oyefeso and colleagues (2000) suggest that, where possible, general practitioners and community health teams should supervise the consumption of medication, particularly in patients identified as high risk. Furthermore, to ensure that drugs provided at any one time do not exceed the patient's therapeutic requirement, GPs and community health teams should be extremely vigilant and study the patient's prescription history, ensuring that patients return unused prescriptions when a change in medication is considered.

Suicide

3.97 It is generally acknowledged that drug overdose is a common method of suicide. Data published by the General Register Office for Scotland on drug-related deaths revealed that, from 1996 to 2006, the number of drug-deaths coded as "intentional self-poisoning" was somewhere between 8% and 18% of the total each year. Darke and Ross (2002) stated that suicide is a major contributor to the fact that heroin users are approximately 13 times more likely to die than their peers and, therefore, a key problem for drug treatment agencies. Farrell and colleagues (1996) highlighted the difficulties in distinguishing between accidental and intentional overdose, in particular among opiate users, and Rossow and Lauritzen (1999) stated that, although the terms "suicide attempt" and "overdose" are theoretically viewed as separate categories, empirically the two types of behaviour cannot always be differentiated. Farrell and colleagues (1996) supported this view stating that the practical management of suicide risk and overdose risk are not very different. However, Darke and Ross (2001) disagree and believe that heroin overdose and suicide present different clinical problems and require different responses.

3.98 Several authors agree on the principal risk factors for suicide. Darke and Ross (2002) underlined that key risk factors for suicide among the general population - such as gender, psychopathology, family dysfunction and social isolation - also apply to opiate users; however, the main issue for heroin users is that they are more widely exposed to these factors. In addition, Darke and Ross (2002) maintain that heroin users carry additional risks associated with their drug use, such as polydrug use, which is linked to an increased risk of suicide. The authors found that drugs play a more significant role in suicide among heroin users than the general population, especially drugs other than heroin, such as benzodiazepines and antidepressants. Farrell and colleagues (1996) also identified the use of several different substances, including opiates, alcohol and benzodiazepines, as an overriding feature of fatal overdose and argue that polydrug use is more likely to be associated with more severe social and psychiatric problems. Consequently, they contend the need for adequate treatment of associated alcohol and benzodiazepines dependence and appropriate treatment of associated psychiatric morbidity in order to reduce the risk of suicide.

3.99 Oyefeso and colleagues (1999) identified the emerging role of antidepressants and the dangers of GPs over-prescribing, in a suicide trends study among UK notified addicts over a period of 25 years. The authors demonstrated the impact that drug control and prescribing attitudes and practices can have on suicide prevention. They refer to how a concerted response from regulatory and treatment sectors in the 1980s led to a substantial decline in barbiturate overdose. Therefore, they contend that an appropriate quantity of antidepressants should only be prescribed after a clear diagnosis of depression and call for closer working relationships between general practitioners and community health teams (Oyefeso et al, 1999).

3.100 In addition, Oyefeso and colleagues (1999) identified methadone as a major contributor to overdose suicide during the last 5 years of the study and, therefore, the need for strict regulation of methadone prescribing and dispensing and supervised consumption, in particular among patients at a high risk of suicide. This is supported by the findings of a study conducted by Darke and Ross (2001) examining the relationship between suicide and heroin overdose among methadone maintenance patients in Sydney, Australia. They found that a history of attempted suicide is common among methadone maintenance patients, in particular female patients, and therefore a major clinical issue for methadone maintenance providers.

3.101 Several papers discuss the possible motivational factors for suicide among problem drug users. One causal factor identified was a distressing life event. Darke and Ross (2001) found that 80% of the sample reported that a major life event, such as imprisonment, had occurred prior to the suicide attempt. Farrell and colleagues (1996) identified loss events, such as loss of a loved one or a job, as a contributing factor to suicidal overdose. Neale and colleagues (2000) agreed that intentional overdoses are motivated by a range of psychosocial factors, such as predisposing personal circumstances and precipitating events, such as arguments, relationship breakdowns and homelessness.

3.102 The findings of the analysis conducted by Neale and colleagues (2000) showed that, consistent with the other papers examined, non-fatal illicit drug overdose is often motivated by suicidal intent. Despite its limitations, a Glasgow study conducted by Jones and colleagues (2002) provided new insight into the extent of contact that problem drug users have with services in the weeks and months before death. They found that 89% of the study sample, for which medical records were available, had seen their doctor in the year before death, often several times, and 20% had been seen by a psychiatrist, who diagnosed suicidal ideation, depression or an anxiety disorder.

3.103 The Glasgow findings are consistent with data from the National Investigation into Drug Related Deaths in Scotland, which revealed that, of the 305 cases for whom records were available, 77% had had contact with general practitioners in the 6 months prior to death, and 17% had had contact with psychiatric services.

3.104 Jones and colleagues (2002) concluded that problem drug users expressing suicidal ideation should be considered at high risk of overdose. An earlier study by Darke and Ross (2001) reinforces this finding as they identified that a quarter of the 223 patients interviewed had severe to extreme depression - almost half reported current suicidal ideation and 61% expressed some degree of hopelessness about the future. Both studies highlighted the significance of this finding for clinical practice as a predictor of future suicidal behaviour. Darke and Ross (2001) suggested the need for careful screening and Jones and colleagues (2002) stressed the importance of a thorough risk assessment by specialist drug services followed up with intensive support. Oyefeso and colleagues (1999) argued that it would be beneficial for primary health care staff, community mental health teams and specialist substance misuse services to carry out suicide risk assessments as part of the routine assessment of problem drug users in their care. Jones and colleagues (2002) described how a specialist co-morbidity team was established in Glasgow in 2000 with the aim of addressing the problems of those with coexisting drug dependence and mental health problems.

3.105 The results of a Norwegian national study of people with drug problems in treatment between 1992 and 1993 identified that engagement in different types of life-threatening behaviour often results from a state of carelessness about life and reflects feelings of indifference, hopelessness and poor self-esteem, which are common due to the chaotic lifestyle and living conditions of drug addicts (Rossow and Lauritzen, 1999). These feelings of indifference and carelessness were identified among a sample of overdose survivors interviewed as part of the National Investigation into Drug Related Deaths in Scotland who attributed their survival to "luck", "God" or having been found on time. Rossow and Lauritzen (1999) highlight that this "indifference" presents a major challenge in providing adequate treatment and support services, which must combine professional competence in substance abuse treatment and psychiatric treatment.

Understanding Social Networks

3.106 It has been suggested that social networks may be an important factor when considering non-fatal overdose. Latkin and colleagues (2004) have suggested that there could be drug overdose links to a) the number of drug injectors in a person's social network and b) the number of networks they might be in conflict with e.g. arguments over sharing drugs.

A) Networks: having more drug injectors in your network can offer more chance to use drugs thus increasing the risk of overdose. Those involved with bigger networks may have a greater number of unplanned injection episodes with less ability to regulate and control their drug use compared to those involved with smaller networks.

B) Conflict: the harmful role that conflict may play within networks was uncovered. Recent overdose casualties reported experiencing conflict with more network members compared to those who had never overdosed or had overdosed in the past (more than 2 years ago). The authors suggest that more severely drug dependent subjects that reported conflictive ties may have been more likely to quickly inject their drug ("slam") because of mistrust of others present within the network.

3.107 An earlier study examining the social context of those who had overdosed found that casualties were more likely to know people infected with Hepatitis C, and shared and received injecting equipment more often with someone who was not their sexual partner (Bennett et al, 1999). According to the authors, overdose casualties are more severely involved in injecting heroin and spend time with others displaying similar characteristics thus normalising their behaviours. The authors suggest that there may be value in trying to influence these cultural norms through peer training and education strategies.

Public Injecting and Overdose

3.108 Injecting drug use in public places is strongly associated with increased risk of blood borne virus transmission, abscesses and overdose (Taylor, 2006). Evidence from research showed that 42% of a sample of needle exchange users had injected in a public area at least once in the week prior to interview (Independent Working Group on Drug Consumption Rooms, 2006). Despite the risks associated with outdoor injecting, heroin use among rough sleepers can be used as a distraction from the discomforts of rough sleeping and a self-medicating means of responding to insomnia and cold weather. However, outdoor injecting under these conditions could contribute to fatal heroin overdoses among rough sleepers through a process of hypothermia secondary to heroin-induced coma (Wright et al, 2005). Where drug use is conducted in street locations, the perception of risk needs to be understood in terms of the social and environmental context in which drug use occurs. Safety from public and police view may be prioritised over the risk of overdose.

3.109 In Australia, Fitzgerald and colleagues (2000) suggested that one factor which could contribute to increased overdose morbidity and mortality is changes in police activity whereby dealing and use at static sites, such as houses, is displaced to street dealing in other areas. Added to this is the unpredictability of heroin quality when scoring in a street environment. It is also suggested that "sensationalist" media coverage of drug use was directly linked to increases in use by acting as publicity for street heroin locations. However in a later study, they described policing strategies that displace public injecting into "quasi-supervised" settings, such as public toilets - providing a degree of "independent third party" supervision that can respond to overdoses (Fitzgerald, 2004). Injecting in shallow or exposed settings increases the risk of discovery by police or public but can also confer a degree of safety from other risks such as overdose or drug-related crime. A recent study on the impact of public injecting highlighted that the public intervened to help overdosed or unconscious drug users, some did this as part of their job, and others did it voluntarily (Taylor et al, 2006).

3.110 Most overdose prevention strategies have focused on changing behaviours. However, the wider context in which heroin use occurs in public environments must be acknowledged and "safer" messages must recognise that "safe" will mean different things to different users depending on their social context. Drug users themselves weigh up the competing risks of public or semi-public sites and future interventions to prevent heroin-related deaths needs to take account of this and should be targeted towards situations where risk is highest.

Accommodation

3.111 Accommodation problems including homelessness have been identified (Neale and Robertson, 2005; Fischer et al, 2004; Wright et al, 2005). For instance, Fischer and colleagues (2004) have identified the important role of housing and other social factors in determining the health of marginalised populations, such as drug users. With housing and other forms of social support having a role to play in reducing drug users' health risks, including the risk of overdose, Neale and Robertson (2005) suggest service providers should seek to identify and address users' problems as part of a broad strategy of overdose prevention - assessment tools could cover a range of personal and social problems.

3.112 Exploring the relationship between housing status, social networks and risk factors for heroin related death, Wright and colleagues (2005) found that various cultures can exist within different types of homeless accommodation that can impact on heroin use. For example, the amount of heroin used, the likelihood of injecting alone or abstaining from drugs. Hostel accommodation appeared to be conducive to group drug use with associated peer pressure, relaxation and availability of heroin and injecting equipment. For some, it was also a place for initiation into injecting heroin use. The hostel setting was also described as a difficult location to stop heroin use due to exposure to triggers for drug use. Paradoxically, although the hostel setting could contribute to one risk factor (increased heroin consumption), the practice of using in a group could also protect against fatal overdoses due to the presence of a third party who could attempt resuscitation and/or alert emergency services. However the presence of a third party could not be viewed as a panacea for all heroin-related deaths as fear of police involvement were cited as reasons for not taking action.

3.113 Looking at rented (social) accommodation, Wright and colleagues (2005) found consistent accounts of young drug users (or those in the early stages of their drug "career") engaging in group drug using activities within the accommodation. Nevertheless, obtaining a tenancy could also increase the potential for fatal overdose due to solitary drug use, a practice that tended to be related to those with a longer history of injecting drug use (Wright et al, 2005).

3.114 Policy implications raised by Wright and colleagues (2005) include the potential for health promotion interventions to reduce fatal overdoses, such as training drug users in resuscitation techniques or in the peer use of naloxone. The authors suggest this would be most effective among those engaged in high risk behaviours, for example injecting with a third party present and also homeless people living in hostel accommodation or using their friends' flats.

3.115 With inherent risks for homeless people engaged in heroin use varying according to their social settings and accommodation, it is suggested here that future overdose prevention initiatives take account of this and target those in high risk situations.

Implications of the Literature Review

3.116 The main implications that can be drawn from the review of literature are as follows:

Emergency responses

• Witnesses present at an overdose event are willing to intervene but the motivating factors that influence intervention and seeking help are complex. They can be shaped by past overdose experience and contact with emergency services. Enhancing effective response may be achieved by offering witnesses (peers, family and friends) a range of CPR training and interventions, such as naloxone or emphasising the need to remain with the casualty until medical help arrives.
• Ambulance data has been used in other countries to identify overdose clusters, those repeatedly overdosing and to map out high risk areas. The overdose data has also been used to offer help to those not in contact with drug services. The lack of UK data on fatal and non-fatal drug overdoses is considered a "significant weakness" in the evidence base thus leading to a call for collating standardised ambulance call-out statistics.
• Reducing "fear" of police involvement through police liaising with ambulance services may increase willingness among overdose witnesses to seek emergency help. Some UK police force areas have developed protocols that avoid police attendance at "routine" overdose incidents by limiting their attendance to fatalities, child protection concerns and threats of violence. Developing protocol changes will require an awareness-raising campaign that disseminates the information to drug users.
• Accident and emergency staff can capitalise on their contact with drug users following an overdose event or other high-risk behaviours (e.g. injecting-related health damage) by offering information on overdose prevention strategies and onward drug treatment referral. Innovative practice is being developed to meet these challenges by locating substance use specialist nurses within accident and emergency departments.

Mainstream & Emerging Interventions

• Reducing drug tolerance risks among those entering/returning to treatment can be achieved through accurate assessment and testing users' tolerance through a process that is reviewed. With many treatment episodes suddenly or prematurely terminated, there is a need to address the potential anti-therapeutic effects of treatment. Improved retention rates and reduced illicit drug use may be achieved by offering most individuals on maintenance methadone a daily dose between 60mg and 120mg (in Britain, only one in 4 service users received over 60mg, according to the NTA).
• Actively involving drug users in their treatment decisions and alternatives to automatic discharge due to non-compliance with treatment regulations should be considered. For example, follow-up assessment of disengaged drug users to increase their uptake and retention within suitable services. It is also important to ensure that short-term prisoners are offered follow-up support from community drug services following release from prison.
• GPs and other primary care staff have a vital role to play in screening for overdose risk factors and provide relevant support. Older heroin users, at risk of overdose, may face a progressive disease burden, thus benefiting from regular health screening and liver function tests.
• There is also a need for clear and enhanced communication between primary and secondary care services involved in prescribing. Key prescribing risk areas are the dangers of "doctor shopping" among drug users, alcohol screening, identification and treatment options and consideration of psychological responses to treat depression, such as counselling, as alternatives to antidepressants. Suicide-risk assessments should also be carried out as part of routine assessments of drug users seeking treatment.
• There is an emerging consensus among the reviewed papers that there is a potential to prevent many opiate overdose deaths using take-home-naloxone, THN. The possible benefits of THN are considered sufficient to justify the need for carefully monitored pilot schemes that are linked into extensive educational programmes and training.
• Safer injecting rooms ( SIRs) may help reduce drug-related deaths, however, the impact of SIRs will depends on factors such as the extent to which it reaches its target population (e.g. homeless drug users) and the number of deaths occurring outside the target population (e.g. drug users injecting at home or socially integrated users). There is also no evidence that SIRs contribute to increased morbidity/mortality risk - no fatal overdoses have occurred within a SIR despite there being "millions" of supervised drug consumptions and thousands of treated emergencies, thus showing evidence that they provide a high level of safety from overdose among people using them.
• Service providers need to be aware of the inherent risks for homeless heroin users which may vary according to their social networks and accommodation. Overdose prevention initiatives need to take account of this and target those in high risk situations.

Developing and Disseminating Key Messages

• Targeted campaigns aimed at addressing key risk factors facing heroin users, such as polydrug use and alcohol consumption, should be considered. These campaigns could also stress the important protective role played by treatment. Disseminating key health messages and overdose prevention information among peer networks may also be an effective way of supporting those at risk.
• It is important to ensure that short-term prisoners receive pre-release information about the risks of overdose. Further information on the risks of cocaine overdose and effective responses is required to increase the understanding and awareness of drug users, peers, family/friends and service providers.

Examples of Innovative Practice

3.117 Some evidence of current practice aimed at reducing drug overdoses was collected. Below are some of examples of innovative practice that have been carried out or which are currently underway in Scotland and England. This is not meant to be an exhaustive list and as such it is accepted that wider activity than is documented in this report might currently be underway. There is limited evidence of review and evaluation of these initiatives, therefore, the term innovative practice has been used throughout this section rather than good practice.

Drug Action Teams

3.118 Several areas of Scotland have set up Action Teams to consider and implement local strategies (e.g. Critical Incident Groups or similar mechanisms) to reduce drug-related deaths. However, there is currently no evaluated evidence to show that these practices reduce drug-related deaths.

Provision of Information

3.119 In 2003 Brighton & Hove DAAT distributed 3,000 copies of a leaflet on overdose and emergency calls. The leaflet provided information for drug users about what would happen if they called the emergency services; why the police might attend; the treatment provided by the ambulance crew; the use of crack and other stimulant drugs; and the recovery position. It also provided useful telephone numbers for services and overdose aid training. The leaflet has since been updated and redistributed.

Training

3.120 For several years now, Brighton & Hove DAAT have commissioned the St John Ambulance Homeless Service to provide overdose aid training to users, family members and friends.

3.121 As part of an effort to reduce drug-related deaths in Greater Manchester, the North West Ambulance Service has been involved in rolling out knowledge of the recovery position.

3.122 Overdose response training, including accredited basic life support training, is provided to service users and carers in Torquay. The Scottish Government has funded a Critical Incidents National Training Officer for almost four years. The National Training Officer provides a range of training initiatives on overdose prevention, primarily targeting service users and providers.

Harm-Reduction Projects

3.123 Set within a traditional needle exchange service, the NHS Lothian Harm Reduction Team, Low Threshold Methadone Programme ( LTMP) was developed to target drug injectors who showed some motivation to change, but found it difficult to keep appointments with mainstream services, and were at risk of overdose due to injecting practice. The LTMP is a self-referral programme that offers a flexible approach to treatment but requires daily attendance for methadone dispensing. There is a key-work system with access to medical, psychological and dental support and BBV interventions. The LMTP team also monitor overdose risk - made easier due to daily contact - with injecting activity recorded and discussed with the keyworker and medical staff. Psychiatric assessment is available to those displaying low mood or express suicidal tendencies.

3.124 HIT, an organisation set up in Merseyside in 1985 with the aim of reducing drug-related harm, currently delivers interventions on drugs, community safety and other public health concerns. Below are descriptions of two of the projects it has carried out to reduce drug-related deaths:

• Lifeguard: Act Fast Save a Life
  • This was a multi-component, social marketing campaign launched in 2003 to reduce opiate-related overdoses based on a collaborative approach across Cheshire and Merseyside. It was commissioned by Cheshire and Merseyside Drug and Alcohol Action Teams, Cheshire Constabulary, Merseyside Police and Mersey Regional Ambulance, NHS Trust. The campaign was aimed at three target groups: opiate users, the general public (including family and friends of opiate users), and practitioners, urging them to "Act Fast, Save a Life" by calling an ambulance at the first sign of overdose. The campaign involved local capacity building, a mass media campaign, and training for professionals and drug users. It also endeavoured to publicise the policy of the police not automatically attending drug overdose incidents when emergency medical help is requested unless exceptional or specific circumstances are identified, such as a threat of violence or evidence of harm being caused to children. The evaluation report for this project is available at:
  • http://www.hit.org.uk/dbimgs/Evaluation%20Report1.pdf
• Peer-To-Peer Project
  • This was a training programme for drug users to challenge misinformation and increase awareness and knowledge of safe practice. The project was created as a result of research showing that injecting drug users are mainly initiated into this practice by their peers and, consequently, it was hoped that by improving participants' knowledge of safe drug using practices; this would in turn improve their confidence and ability to pass on this knowledge to the wider drug using community.

Naloxone Pilots

3.125 Several naloxone pilots have already taken place in Scotland and England. In Scotland, these pilots have been carried out in Lanarkshire and Glasgow.

3.126 The Lanarkshire Naloxone (Narcan ^®) Pilot was designed to provide users, their family and friends, and service providers with another mechanism in overdose management aimed at reducing drug-related death by training and educating them in basic life support, the treatment of the unconscious patient and the administration of naloxone. Those running the pilot also wanted to assess whether it was possible to deliver an effective training programme covering critical incident management, and safe and effective naloxone administration, and whether clients could demonstrate responsible management of naloxone and effective use in an overdose situation. So far, approximately 42 people have been trained as a result of this pilot and 34 take-home naloxone kits distributed. Provisional results show 2 successful "saves" by pilot participants and no inappropriate use of naloxone has been reported.

3.127 The Glasgow Naloxone (Narcan ^®) Pilot was conducted in a similar manner over a 1-year period with service users, their families and concerned others. It involved providing family members, carers and service users with their own personal supply of take-home naloxone. As with the naloxone pilot in Lanarkshire, all participants were given training and information on basic life support techniques, overdose awareness (i.e. how to recognise the symptoms and risk factors), and how to administer naloxone safely and responsibly. Take-home naloxone was initially supplied to approximately 250 service users through the Glasgow Drug Crisis Centre. An evaluation of this pilot was carried out involving 2 separate questionnaires: one for drug users and another for their family and carers. The results showed that 251 supplies of take-home naloxone were provided between April 2007 and March 2008, with 12 reported appropriate uses of naloxone in an overdose situation.

3.128 In April 2008 the Medical Research Council approved funding for a UK wide prison research study. The study aims to measure the success of providing naloxone on release from prison in preventing heroin-related overdose. The Research Team have been planning the randomised control study, in consultation with Scottish Prison Service ( SPS) and Her Majesty's Prison Service ( HMPS), for 2 years prior to being awarded funding. Although SPS staff may be involved in prisoner training and distribution of naloxone packs on release, there is no financial commitment from SPS required. Roles and responsibilities of both SPS and the Research Team are being clarified and preparatory work must be completed before any research begins later this year.

3.129 In terms of England, Salford DAT and North West Ambulance Service were involved in a national Take-Home Naloxone Project in 2006, which was rolled out to all service users in Tier 3 throughout 2007. A training video was produced as part of this project. Wiltshire DAAT was also involved in a Naloxone Project, run by a service user forum, which involved overdose training for service users and carers.

Ambulance Protocol

3.130 A similar protocol regarding police attendance at overdose incidents to that publicised as part of the Lifeguard Project in Cheshire and Merseyside was introduced in Nottinghamshire in 2000. An agreement was reached between Nottinghamshire Police, the East Midlands Ambulance Service and the local DAATs to ensure that police officers do not routinely attend ambulance call-outs to drug overdoses unless a death has already occurred; there are child protection concerns; and/or the address is identified as one where there could be a threat of violence. Similar protocols have also been established in other parts of England, such as Kirklees, Leicestershire and Avon & Somerset.

Drug-Related Death Partnership

3.131 The Drug-Related Death Partnership ( DRD Partnership) is a multi-agency partnership set up in Oxfordshire with the aim of reducing the number of drug-related deaths in this area. The Partnership produced a Drug-Related Death Strategy for 2006 to 2009 in accordance with the National Treatment Agency's ( NTA) treatment effectiveness agenda and national programme in England to reduce drug-related deaths.

3.132 The strategy document contained seven strategic objectives for this area and provided a summary of a number of projects and protocols that have been established by, and inform the work of, the DRD Partnership:

• Communication about Acute Risks
  • This involves an Early Warning System, co-ordinated by Oxfordshire DAAT, to rapidly inform drug users, carers and staff about strong or adulterated drugs in local circulation.
  • It also includes a Communication Protocol to ensure accurate and meaningful information is provided to the public about acute risks of illicit drugs locally. For example, Oxfordshire DAAT, in collaboration with the appropriate communication departments in Oxfordshire's healthcare system, currently co-ordinate the dissemination of warning messages and other information through the local media.
• Police Attendance at Overdose Incidents
  • To encourage drug users to contact the emergency services in case of overdose, Thames Valley Police, Oxfordshire DAAT and Oxfordshire Ambulance Service reached an agreement in March 2004 that police will not routinely attend overdose incidents in Oxfordshire.
• Overdose Prevention and Response Training
  • Oxfordshire User Team and Oxfordshire Ambulance Service have been delivering training on overdose prevention and response to drug users and their carers since 2002. The aim of a series of training workshops being carried out is to improve communication between drug users and the emergency services, provide up-to-date information on risk, and enable users and carers to practice basic life support skills.
• Confidential Inquiries into Drug-Related Deaths
  • The DRD Partnership reviews every drug-related death that occurs in Oxfordshire to identify risk factors. The findings from these Confidential Inquiries are used to inform service provision, improve interventions and reduce potential risks.
• Publications
  • Oxfordshire User Team has produced local publications on overdose prevention and response, safer injecting techniques and hepatitis C.
• Police Trained to use Breathing Apparatus
  • Police working in Oxfordshire have been trained by Oxfordshire Ambulance Service to use breathing apparatus (bag/valve/mask), and breathing apparatus is now kept in police cars.

Joint Working Practices

3.133 Bennett and colleagues (2006) described some examples of good joint working practice introduced in the Brighton & Hove area between 1998 and 2006 in response to the recommendations arising from a Confidential Inquiry into the high drug-death rate in this area. Some examples of innovative co-ordinated practice in Brighton & Hove include:

• Front line ambulance staff are rotated through the substance misuse harm reduction clinic.
• A joint assessment of Drug Treatment and Testing Order ( DTTO)/Drug Rehabilitation Requirement ( DRR) clients is carried out on the same morning by probation, health and voluntary sector providers.
• The local NHS substance misuse service is responsible for managing the prison substance misuse team.
• The A&E Department substance misuse nurses report on the previous day's admissions and attendances on a daily basis in order to enable plans for hospital discharges and prison releases to be made.
• A nurse has been included on the arrest referral team.

Festive Overdose Awareness Campaigns

3.134 A representative from SPSHQ Addiction Team is a member of the Preventing Overdose Campaign Group in Glasgow. SPS is actively involved in the Festive Overdose Awareness Campaign, ensuring that individuals released over the festive period (between December and January) are offered overdose awareness information using the same methods issued by this group to community services for that year (key rings, red information cards and facemasks). This is carried out within all prisons not just those who release individuals to the Glasgow area. SPS also displays posters promoting the Preventing Overdose Awareness Event held in Glasgow and encourages those being released to the Glasgow area to attend.

Harm Reduction Measures in SPS

3.135 Since October 2005, SPS have provided a Needle and Syringe Pack to prisoners leaving custody (and storing them on entry) in 6 prisons throughout Scotland ( HMP Aberdeen, HMP Barlinnie, HMP & YOI Cornton Vale, HMP Dumfries, HMP Glenochil and HMYOI Polmont) in support of the schemes operating in Police Custody Suites. This was piloted in HMP Aberdeen and rolled out to 4 additional prisons. HMP Barlinnie began providing these packs in early 2006 and HMP Edinburgh is currently working to implement this measure.

3.136 In November 2005 HMP Aberdeen piloted the provision of paraphernalia to injecting drug users ( IDU) in custody. The pilot ran for 18 months and IDU prisoners were provided with water for injection, citric acid, spoons, filters, pre- and post-injection swabs, information leaflets on local services and the use of equipment, and one-to-one sessions on safer injecting with an addictions nurse; however, needles and syringes were not provided. This measure was re-named the Harm Reduction Protocol and had been rolled out across the entire prison estate by March 2008. Needle and syringe provision to IDU prisoners was recommended as a public health measure in July 2005.