Why are drug-related deaths among women increasing in Scotland? - full report

1. Background

1.1 Introduction

In Scotland, drug-related deaths are reported on an annual basis by National Records of Scotland ( NRS), using a definition introduced in 2001 (known as the NRS implementation of the ‘baseline’ definition for the UK Drugs Strategy). This definition includes deaths attributed to the following causes (with ICD-10 codes given in brackets):

• deaths where the underlying cause of death has been coded to the following sub-categories of ‘mental and behavioural disorders due to psychoactive substance use’:
  • opioids (F11);
  • cannabinoids (F12);
  • sedatives or hypnotics (F13);
  • cocaine (F14);
  • other stimulants, including caffeine (F15);
  • hallucinogens (F16); and
  • multiple drug use and use of other psychoactive substances (F19).

• deaths coded to the following categories and where a drug listed under the Misuse of Drugs Act (1971) was known to be present in the body at the time of death (even if the pathologist did not consider the drug to have had any direct contribution to the death):
  • accidental poisoning (X40 – X44);
  • intentional self-poisoning by drugs, medicaments and biological substances (X60 – X64);
  • assault by drugs, medicaments and biological substances (X85); and
  • event of undetermined intent, poisoning (Y10 – Y14).

Additional information on the definition of drug-related deaths can be found in Annex A of the most recent NRS report (National Records of Scotland, 2017a).

Drug-related deaths have been increasing in Scotland in recent years (National Records of Scotland, 2017a). 2016 saw the highest number of deaths ever recorded, at 867 – a 106% increase on the figure from 2006. Although men still account for the majority of these deaths, the proportion of women has increased over time, from 19% in 2002-2006 to 29% in 2012-2016. When comparing the annual average for 2012-2016 with that for 2002-2006, the percentage increase in the number of DRDs was greater for women (169%) than for men (60%). The increase in deaths among women over the last decade has also been more consistent, with less year-on-year variability than among men despite smaller numbers. Figure 1 shows the overall trend in number of deaths by gender.

A similar trend has also been observed over the same period in England and Wales, though appears to be somewhat less marked ^[1] . Between 2002-2006 and 2012-2016, the average annual number of drug misuse deaths among women increased by 64%, compared to 28% for men (Office for National Statistics, 2017).

Figure 1. Number of drug-related deaths in Scotland 1996-2016, by gender.

Source: National Records for Scotland

1.2 Aim

In response to these figures, a rapid scoping project was undertaken to examine potential explanations for the rising rate of DRDs among women in Scotland in recent years and to identify implications for policy.

This paper summarises the findings of that project. It starts by describing the methods used and providing a general overview of gender and drug-related harms, before considering a range of possible explanations in turn. It concludes with a summary of potential implications for policy, practice, and research.

1.3 Methods

This project consisted of four strands:

• Review of relevant routine data sources, as described below. All data not routinely published but obtained through bespoke requests were provided in aggregate form for the purposes of confidentiality.
  • Drug-related death registrations, National Records for Scotland
  • National Drug-Related Deaths Database, Information Services Division Scotland (ISD) – in particular, the report published in 2016 covering deaths between 2009-2014 which included a specific section on gender ^[2]
  • Scottish Drugs Misuse Database ( SDMD), Information Services Division Scotland ( ISD)
  • Drug-Related Hospital Stays data, Information Services Division Scotland ( ISD)
  • Needle Exchange Surveillance Initiative ( NESI), Health Protection Scotland/University of West of Scotland/Glasgow Caledonian University
• An overview of relevant literature, based on a search for existing systematic reviews undertaken by Scottish Government library services (details of which are included in appendix 1), ‘snowballing’ of reference lists to identify relevant articles and reports, and follow-up of specific sources mentioned by key informants. Both peer-reviewed and grey literature was considered for inclusion, though priority was given to studies undertaken in the UK, Europe and other high-income countries. No date restrictions were imposed. Findings from the literature were coded using a thematic matrix reflecting existing and emerging explanations. Due to time constraints, a comprehensive literature search or quality assessment of individual papers could not be undertaken, so the overview of the literature reported here is not equivalent to a formal review. Very few relevant systematic reviews were identified, so there was greater reliance on non-systematic reviews and snowballing. These limitations should be borne in mind when interpreting the results.
• Engagement with key stakeholders from statutory services, the third sector, and academia through fifteen interviews with 16 participants; two focus groups; and one observed meeting of an existing community of practice. These were identified through existing networks such as the Partnership for Action on Drugs in Scotland Harm Reduction Group and local Alcohol and Drug Partnerships, as well as professional contacts of the author and policy colleagues. Potential participants were invited to participate by email and provided with an information sheet at the point of invitation; verbal informed consent to participate was obtained prior to embarking upon the interview. A summary of informant roles and an example topic guide is provided in appendices 2 and 3. Interviews (either in person or by telephone) and focus groups were undertaken by an experienced qualitative researcher, with notes transcribed immediately after the event. Emerging themes were identified by collaborative review of notes between the authors, using the thematic matrix described above. The matrix was iterated to reflect additional themes as they emerged. As these interviews and focus groups were recorded through written notes rather than audio tape, no direct quotations are included in this report.

• Analysis of the transcripts of 28 semi-structured interviews with women over 35 years of age who use drugs, initially undertaken as part of the Older People with Drug Problems ( OPDP) project (Matheson et al., 2017). This secondary data analysis approach was chosen to avoid imposing additional research burden on potential participants, particularly given the sensitivity of the topics at hand. Participants in the OPDP were recruited through a range of non -NHS settings across Scotland (such as needle exchanges, counselling, voluntary organisations, and homelessness services), using a quota sampling approach based on the age and gender profile of the Scottish drug using population. Interviews were undertaken by trained peer researchers, all of whom happened to be women: signed consent was obtained from each participant. For this project, transcripts of interviews with 28 of the 30 women in the sample were selected and coded by an experienced qualitative researcher using the thematic matrix described above, using directed content analysis. Again, the matrix was iterated to reflect additional themes as they emerged. Quotations from these transcripts are included throughout the report to illustrate key themes, alongside the (non-consecutively numbered) interviewee code. All names of people, services, and locations have been anonymised.

1.4 A note on terminology

The terminology of sex and gender have subtly different meanings: sex (females/males) refers to biological characteristics, whereas gender (women/men) refers to self-perception and socially constructed role.

However, these terms are often treated as interchangeable, and are rarely used consistently. Because of this lack of consistency, and the multiple sources on which we draw, in this report we have defaulted to using gender and women/men (though we continue to use female/male as adjectives). This decision reflects a necessary pragmatism, rather than an explicit theoretical or empirical orientation.