The Scottish Government

« Previous | Contents | Next »

Listen

Chapter 1: Introduction

What is in this chapter? This chapter summarises the information that is already known about drug related deaths. It describes the background to the commissioning of this study, sets out the aims and objectives and describes the methods employed to meet these.

Where did the information come from? The background section is drawn from published international research.

Purpose of this report

In 2003, the General Register Office for Scotland (GROS) reported the country's highest ever annual number of drug related deaths, 382 deaths during 2002 (GROS 2003). The Scottish Deputy Justice Minister subsequently called for a national investigation into all drug related deaths in Scotland for the year 2003. This report is the outcome of that investigation.

A great deal is already known about the circumstances and toxicological findings of drug related deaths (DRDs), especially those involving heroin and methadone. There is now a strong research evidence base indicating that many of these deaths are preventable. However, less is known about the social and clinical circumstances of these individuals prior to their deaths, in particular the nature of contact made with treatment agencies, prison service and other services. This study explores the patterns of contact with service providers in the 6 months prior to death, and examines implications of these findings for services. It provides the most recent information to date on the toxicology and circumstances of death of drug-related deaths in Scotland, and compares these findings with data emerging from a concurrent study into drug-related deaths in London for the same year. The opinions, beliefs and knowledge of a group of living injectors who have survived a recent overdose are also reported here.

Background

Known characteristics of drug related deaths

Remarkably, the characteristics, correlates and contexts of opioid drug related deaths examined in different countries and at different times are consistent (Best et al 2000). Consistent findings include:

• High prevalence of poly-drug use, especially benzodiazepines and alcohol (Darke and Zador 1996, Risser et al 2000, Oliver and Keen 2003)
• Loss of tolerance due to recent abstinence from heroin e.g. persons recently released from prison or abstinence based inpatient programmes (Seaman et al 1998, Bird and Hutchinson 2003, Strang et al 2003)
• Deaths occurring in the presence of others and/or in a home environment. (Zador et al 1996, Darke et al 2000, np- SAD 2002)

All of these circumstantial factors strongly imply the opportunity for intervention prior to death.

White and Irvine (1999) have speculated that resistance to respiratory depression may be less complete and fall more rapidly than resistance to the 'euphoric' effects of heroin. Thus, as injectors consume more drugs the gap between euphoria and overdose narrows, and any 'dip' in tolerance due to interruptions in drug use precipitated by imprisonment, treatment or unfamiliar environment may prove fatal (White and Irvine 1999, Siegel 1984).

Route of administration is also predictive of increased mortality. Most heroin related deaths have occurred following an intravenous injection (Darke et al 2000). Overdose events were rare among smokers of heroin (Gossop et al 1996). Oliver and Keen (2003) recently confirmed these findings when they reported that the intravenous route of administration was used in 84% of DRDs in Sheffield, England.

Methadone related deaths, as for heroin deaths have been found to be primarily poly-drug deaths, few are positive only for methadone (Perret et al 2000, Sunjic and Zador 1999). An international review of methadone related deaths indicated that the majority were diversion related i.e. occurring in persons not prescribed methadone at time of death (Zador 1999), a finding replicated by the most recent published Scottish investigation of methadone related deaths (Seymour et al 2003).

Further examination of the death certificates and coronial records in the NTORS cohort (Gossop et al 2002) were consistent with previous findings: the majority of deaths were found to be accidental, with more than one drug being detected at autopsy in most cases, occurring in a home environment, and with half of cases dying in the vicinity of another person. Non-prescribed benzodiazepine use, harmful alcohol consumption, anxiety and homelessness were found to be predictive of increased mortality.

Scottish data on all drug related deaths ( GROS 2003) in 2002 indicated that the sociodemographic profile, toxicological findings and circumstances of death had remained largely unchanged. Most cases of DRD continued to occur in males under the age of 45 years, predominantly involved heroin/morphine with a high proportion positive for benzodiazepines and alcohol, and primarily occurred in known or suspected drug abusers.

Long term trends of drug-related mortality and fatal overdose suggest that they may have fallen from over 2% per annum among opiate users recruited from 1968-1976 to 1% in 1986-1992 (Ghodse et al 1998). Two recent record linkage studies in the UK reported that among the NTORS cohort the annual mortality rate was 1.2% after four years follow-up (Gossop et al 2002), and in a pilot cohort recruited in London the annual mortality was 1.6% after an average of two years follow-up, 17 times higher than the mortality rate of the London population (Hickman et al 2003a).

Among Glasgow injectors followed up after entry to methadone treatment in 1996, 1.5% died in the first year (Hutchinson et al 2000), and this annual rate was sustained in the subsequent 4-5 years.

Risser et al (1996) established a profile of the family background of Austrian cases of drug related death based on interviews with family members of the deceased. This study found that cases experienced high rates of traumatic events during childhood, predominantly parental divorce or death, initiated alcohol and tobacco smoking by age 15, and that a high proportion of cases had a family member with a drinking problem, among other findings. This highlights the importance of the social and family contexts of DRDs.

Research into DRDs is now building on the above findings to examine the implications for prevention i.e. interventions to reduce DRDs.

Survivors of drug overdose

It is instructive to note the findings from research to date into survivors of drug overdose. Neale (2000) interviewed 77 resuscitated heroin users in two Scottish accident and emergency departments, and found that attempted suicide was the reason underlying the overdose in 49%. However, interviews with a sample of Australian drug users found that deliberate heroin overdose as a method for attempting suicide was reported by only 10% (Darke and Ross 2001). These authors followed up with an international literature review of rates, risk factors and methods of suicide among heroin users and found that while this group was 14 times more likely than peers to die from suicide, deliberate overdose of heroin as a means of suicide was infrequently used compared with prescription non-opiate medication overdose or more violent methods (Darke and Ross 2002).

Perception of risk

In other research into survivors of drug overdose, Darke and Ross (1997) found that 80% of survivors of heroin overdose did not perceive themselves at high risk of overdose, despite experiencing one in the previous six months and despite research indicating that overdose was a common occurrence in heroin users (Darke et al 1996). Zador et al (2001) confirmed this perception of a low personal likelihood of a subsequent overdose among heroin users interviewed within seven days of their most recent overdose. Furthermore, only 11% stated that they would seek treatment when asked what impact their recent overdose would have on their future drug using behaviour. The majority of individuals indicated that they would continue to use heroin albeit more safely, or try to stop.

Other research has shown that although drug users give many reasons for seeking treatment (desire to stop using drugs, tired of the heroin-using lifestyle, fear of re-incarceration or relationship breakdown), experience of overdose or worry about having one was rarely one of these reasons (Weatherburn and Lind 2001, Sell and Zador 2004).

Recent research has revealed the high prevalence of morbidity in injecting drug users both as a cause and consequence of overdose (Warner-Smith et al 2001) which indicates the need to emphasise the greater range of health consequences of a non-fatal overdose in harm reduction messages to injecting drug users. The offer of intervention e.g. provision of an information card about local agencies and services to heroin users at the time of overdose by ambulance officers has been shown to be successful at targeting an often 'hidden population' of high risk users who might not otherwise be in contact with treatment or other services (Dietze et al 2002).

This evidence base implies that experience of overdose is neither a deterrent to future drug use nor a motivator for seeking treatment. If overdose is personally contextualised as an acceptable "occupational risk" by injecting drug users, along with the risk of blood-borne viral infection and incarceration, then this would imply a different overdose prevention strategy from that where overdose is contextualised as a marker of underlying psychopathology. The complex relationships between experience of drug overdose, perceived risk of overdose, depression and suicide need further exploration. At the very least, research to date suggests that future strategies designed to prevent drug overdose need to target risk perceptions, knowledge of risk factors for fatal overdose, as well as mental health status and other factors within this high risk group. However, these findings need to be replicated in Scottish drug users.

Aims and Objectives of the Study

Aims

1. To collect and analyse information relating to the clinical and social circumstances surrounding all drug-related deaths in Scotland for the period January - December 2003.
2. To identify patterns in social and clinical circumstances surrounding the deaths, and the associations between them.
3. To make recommendations for policy and practice which may lead to a future reduction in drug-related deaths.

Objectives

1. For all individuals: to establish and explore the nature of the individual's social circumstances prior to death including employment and accommodation status, social networks and family circumstances, and the nature of the individual's drug use.
2. For all individuals: to determine the specific combinations of drugs associated with each death: to analyse patterns in the toxicology reports on each death, to determine the involvement of illicitly-obtained prescription medication, to investigate differences between injecting drug users and non-injecting drug users in the circumstances of death.
3. For all individuals: to establish the nature and extent of contact with "services" prior to their death.
4. For those individuals who were in contact with "services" prior to death:

• To determine the nature of the individual's contact. To determine the individuals who had a previous non-fatal overdose, and whether there was any follow-up by "services".
• To determine the individuals who had been released from prison, and whether they had been offered, and had participated in the prison Transitional Care arrangements.
• To determine the nature of the individual's contact with the wider criminal justice system ( e.g., the police and court system).
• To collect detailed data on the care of individuals who had received treatment (including a detoxification programme) for drug misuse, to compare the clinical care received by those who were in treatment with national clinical guidelines, to consider whether there was anything that "services" could have done, or could do in the future, to prevent death.

5. For individuals who were not in contact with services prior to death: to identify whether individuals were unknown to "services" or whether there had been breakdowns in the care pathways for these individuals, and to determine which, if any, were waiting for drug treatment, and explore the contribution of this to their death.
6. On the basis of the information gathered for 1-5 above: to explore the possibility of drawing conclusions about regional differences in deaths, and to determine whether the circumstances associated with 2003 deaths are broadly reflective of what is already known about risk factors from previous research.
7. Investigate patterns in broad demographic variables ( e.g., gender, age, etc.) for the period 1996 - 2003, investigate patterns in gender and age.

In addition, the study includes the following aims:

1. Extended interviews with a sub-sample of living injectors who had survived an overdose in the 6 months prior to interview to obtain more information about survivors' beliefs, experience and knowledge of overdose.
2. A comparison of the characteristics of Scottish drug related deaths with a sample of London drug deaths for the same year (2003).

Methods

(1) Drug related deaths in Scotland, 2003

The study population comprised 317 cases of drug-related death in 2003 identified by GROS. Case definitions of drug related deaths were based on ICD-10 codes for drug-related death categories (see Appendix 1). In brief, each case of a drug related death as found by the Procurator Fiscal was entered by the GROS into one of four main ICD-10 categories: 'mental and behavioural disorders due to psychoactive substance use' (F11, 13, 19), i.e. accidental drug overdoses occurring in persons with drug dependence or abuse, drug-related suicides (X61, 62), 'poisoning by narcotics of undetermined intent' (Y12), and 'accidental poisoning' (X41).

Prior to commencement of the study, full ethical approval was obtained from the MREC Scotland. The study was conducted within the research governance framework for health and community care ( SEHD, 2001) and was externally audited in February 2005. NHS Forth Valley acted as research sponsor for the investigation.

Before contacting NHS services each local research ethics committee ( LREC) was informed of the study and management approval to conduct the study was obtained from each research and development committee.

The Association of Directors of Social Work ( ADSW) standing committee on standards, training and research gave formal support to this study in January 2004. In addition each Director of Service or equivalent was contacted for permission prior to approaching individual social work services.

Two standardised data collection forms were developed: The clinical and social circumstances ( CSC) form was used to record information obtained from case notes held with social care services, specialist drug treatment services, mental health services, and non-statutory agencies across all health board areas in Scotland. The fiscal file ( FF) data collection form recorded data on the circumstances of death, post-mortem toxicological findings and cause of death sourced from deceased cases fiscal files, sent down by Procurators Fiscal offices to the Crown Office.

Copies of these data collection forms may be obtained from the investigators.

In addition, information was collected from the Scottish Prisons Service ( SPS), Scottish Criminal Records Office ( SCRO), Practitioner Services Division ( PSD) and Information and Statistics Division ( ISD) of the Common Services Agency ( SMR 00 (outpatient attendances), 01 (acute hospital admissions) and 04 (psychiatric hospital admissions) databases.

The following variables of information were collected: sociodemographic characteristics including accommodation, employment and relationship status, circumstances of death and post-mortem toxicology findings, cause of death, pattern of contact with health, social and criminal justice agencies in the 6 months prior to death including reason for, frequency of, and outcome of contact, medical and psychiatric history, and drug treatment details, history of non-fatal overdose and relevant details of offending history including incarceration.

A team of six data collectors were recruited to the study and trained in the use of the data forms and in the interpretation of service case note and fiscal file information by three of the investigators ( BK and AB, and DZ respectively). The data collectors were under the management and regular supervision of the research project manager ( AR). They read files and recorded relevant data onto the CSC and FF forms.

A list of all drug services in Scotland was compiled by merging three existing databases: the Corporate Action Plans of the 22 Drug and Alcohol Action Teams, the services database of The Centre for Drug Misuse Research, University of Glasgow and the Scottish Drugs Forum Directory of Specialist Helping Agencies. In addition, all NHS mental health services and local authority social work departments were included.

The project manager made initial contact with each agency normally by phone, outlining the nature of the investigation and asking for agreement in principle to access any records that they might hold. A letter detailing the aims and objectives of the study, relevant permissions and approval and a formal request for access to records followed this up. Once permission had been given the geographic catchment area of each service was established and a list of names of people who died within that area was provided to them. This was normally sent as a password-protected spreadsheet to a named individual within the agency. That person normally chose the password. On occasion, with the prior knowledge and agreement of the named individual, the list of names was posted to an agency addressed to that person and marked 'private and confidential'.

After establishing which people had been known to each service, arrangements were made for the data collectors to visit the agency in order to examine the case records and extract the data required to complete the CSC form as fully as possible. Table 1.1 below sets out the number of agencies contacted in each health board area.

Table 1.1: Number of agencies contacted in each health board area

In total 175 services were contacted, of which 173 (99%) gave their permission for the data collectors to access their records.

Data recorded on the CSC and FF forms as well as selected data provided by ISD were entered into SPSS (v. 12) for quantitative analysis. These analyses were descriptive.

(2) Trends in drug related deaths in Scotland, 1996-2003

Data collected by the General Register Office for Scotland ( GROS) on drug-related deaths in Scotland during 1996-2003 were examined for trends in all deaths and those involving either heroin/morphine or methadone. Poisson regression was used to assess the rate of change in drug-related deaths during 1996-2003 by cause of death, day of death, gender, age, health-board area and drugs specified on the death certificate.

(3) Experiences of overdose survivors

Injecting drug users ( IDUs) were interviewed as part of an ongoing study evaluating the change in the Lord Advocate's guidelines on the supply of needle and syringes to injecting drug users. A multiple site recruitment strategy was used to generate as representative a sample as possible of IDUs in Glasgow. Sites included pharmacy exchanges, needle exchanges, drug treatment services and various street sites throughout Glasgow. Recruitment was undertaken in two phases. The first phase took place between February and April 2004. The second phase began in July 2004 and ended in October 2004.

Criteria for inclusion in the study was having injected a drug in the previous month.

Interviewers explained the study to all respondents and assured them that all information provided by them was anonymous and confidential. Those who had experienced an overdose in the previous 6 months were asked to consent to a follow-up interview and asked to provide a contact number and address. Interviews took place in a venue chosen by the injector (usually their own home) and were audio-taped.

The interviews sought to determine:

• IDUs knowledge of risk factors for overdose
• IDUs overdose risk behaviours
• IDUs knowledge of appropriate interventions in overdose situations
• Why respondents believed they had survived their overdose
• How overdoses could be prevented
• If respondents were in treatment at time of overdose
• If not in treatment, why not
• IDUs personal perception of overdose risk.

Audio-taped interviews were transcribed and analysed by summarising responses and identifying themes. Quantitative data were entered into SPSS (v.12) for analysis.

(4) Comparison of Scottish and London Drug Related Deaths ( DRDs)

The Scottish data were extracted from Procurator Fiscal files as described above. London coronial data were extracted from 7 of the 8 coronial courts, accounting for approximately 75% of drug related deaths in London. As the GROS criteria for identifying cases of drug related deaths were not used in the London study, deaths from both samples were selected that were positive on post-mortem toxicology for morphine (heroin), methadone, cocaine, MDMA, amphetamines, or dihydrocodeine yielding 273 deaths for Scotland and 148 for London. The number of Scottish cases used for this part of the study is lower than the total number of deaths investigated in the main report, as deaths not positive for any of these drugs were excluded.

All data were entered into SPSS (v.12) for quantitative analysis. Simple descriptive analyses assessed differences between London and Scotland in terms of age, gender, toxicology, homelessness, history of imprisonment, drug treatment, place of death, number of witnesses and ambulance attendance.

« Previous | Contents | Next »