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

3. Evaluation of potential explanations

3.1 Introduction

This section attempts to go beyond the relatively static description of gender differences in drug use, risks, and harms in the preceding section to identify factors which might explain the trend in drug-related deaths among women. This is a much more challenging prospect, for a number of reasons: up-to-date evidence on relevant factors in recent years is limited; what is available does not lend itself to formal hypothesis testing; and in reality the causes of this phenomenon are likely to be multiple and interacting.

This section therefore uses the synthesis of material from all three strands to give a narrative summary of the evidence examined, rather than definitive conclusions.

It starts with artefactual explanations, then examines individual, community and relationship factors, and finishes by examining wider social and economic trends and policy context. It is therefore not organised in order of importance or relevance.

The potential explanations examined are:

• Changes in the definition or recording practices for drug-related deaths
• A hypothesised increase in mortality among women in the general population
• Changes in the number of women who use drugs
• Ageing among women who use drugs
• Co-occurring physical and mental health conditions
• Changing patterns of drug use among women
• Changes in relationships and parenting roles
• Experience of adversity, trauma, and violence
• Access to and engagement with treatment and harm reduction services
• Experiences of prison and liberation
• Economic and social trends, including austerity and welfare reform

3.2 Changes in definition or recording practices for drug-related deaths

Analysis of National Records of Scotland ( NRS) data undertaken for this project show that the trend in drug-related deaths among women is not substantially affected by changes in the classification of drugs since 2000, including the classification of tramadol and novel psychoactive substances such as mephedrone (see additional table 1 in Appendix 4).

NRS describe a number of factors which may affect comparisons of drug-related death rates between different years and different areas, including technical advances resulting in more sensitive laboratory tests, changes in testing practice, or pathologist decision-making when identifying substances as implicated in, or potentially contributing to, the cause of death (National Records of Scotland, 2017a). However, these were reported by the forensic toxicologist interviewed to be highly unlikely to differ by gender.

This artefactual explanation is therefore unlikely to explain the trends observed.

3.3 A hypothesised increase in mortality among women of this age in the general population

One query raised in relation to the trend observed is whether it merely reflects background changes in mortality risk or overall health status among women of this age group in the general population.

A review of crude and age-standardised mortality rates for women in the general population of Scotland between 2001 and 2016, supplemented by discussions with the NRS statistician responsible for drug-related deaths data, found that mortality rates in the relevant age bands are generally static or declining (National Records of Scotland, 2017b). Where this was not the case (slight increase in crude mortality among women aged 35-50 years since 2014), it is difficult to separate out the contribution of drug-related causes, which make a substantial contribution to the relatively low numbers of deaths in these strata of the population. In comparison, male all-cause mortality rates in several ‘middle-age’ strata appeared to show more significant increases in recent years.

The general population incidence of, and mortality from, cardiorespiratory conditions which might plausibly interact with drug use to cause the observed trends have also remained static or declined among women in the relevant age groups over the last fifteen years (Information Services Division Scotland, 2017b, Scottish Public Health Observatory, 2017b).

Together these data suggest that changes in background mortality risk or overall health status in the general population are unlikely to explain the trend observed.

3.4 Changes in the number of women who use drugs

Increasing numbers of deaths among women who use drugs might reflect an increase in the number of women who use drugs, the risk of death experienced by that population, or a combination of the two.

Understanding the scale of problem drug use – particularly over time – is challenging. The most recent estimates of the prevalence of problem drug use in Scotland are from 2012/13 and did not report on trends in prevalence separately by gender (Information Services Division Scotland, 2014). Although new estimates of the prevalence of problem drug use in Scotland are currently being prepared, and will include detailed gender breakdowns, the limitations of the data sources (such as lack of linkable identifiers and potential sources of bias) and lack of continuous data mean that trend data for DRD rates per 1,000 people with problem drug use cannot be calculated.

In the meantime, the most robust indications of the size and gender composition of the population at risk of drug-related death come from population surveys and data from treatment services. Each has strengths and limitations but in combination they may give an indication of the overall trend.

The Scottish Crime and Justice Survey has found that the prevalence of self-reported drug use in the past year among women in the general population has fallen in recent years, from 4.3% in 2008/09 to 3.4% in 2014/15: prevalence also declined among men, albeit from a higher baseline (Scottish Government, 2016). The NESI survey of injecting equipment provision clients has found an ageing population with increasing average time since onset of injecting among both genders, consistent with declining recruitment into injecting drug use (Health Protection Scotland, 2017b). The gender profile of the NESI sample has been similar across each two-year sweep (Health Protection Scotland, 2017b).

The number of women undergoing initial assessment for specialist drug treatment in Scotland has remained largely stable over the past decade, reflecting little change in both the overall number and gender profile of people attending for initial assessment ^[3] (Information Services Division Scotland, 2017d). During the last five years, the gender profile of clients attending for initial assessment has been similar to that of cases of drug-related death, at around 29% women. Prior to this, women accounted for a slightly higher proportion of those attending for initial treatment than those dying from drug-related deaths (30% vs 22%) (Information Services Division Scotland, 2017d, National Records of Scotland, 2017a).

Similarly, drug-related hospital stays – which will include some people with problem drug use not accessing specialist drug treatment – have shown little change by gender until around 2013/14, when there was a rise in the rate of new patients, an increasing proportion of whom were men (Information Services Division Scotland, 2017a).

Together these data suggest that the increase in drug-related deaths among women in recent years is less likely to be driven by an increase in the number of women using drugs.

However, given the limitations of population surveys and treatment service data, they do not completely exclude the possibility of ‘hidden’ populations of people who use drugs (particularly prescribed drugs) which may be growing in number and/or experiencing increased risk of death.

For instance, several stakeholders highlighted the dependence on data gathered in treatment services and the limitations of this for understanding the population of people at risk of drug-related death. This may be particularly true for women, who were felt by stakeholders to be less likely to engage with services and less able to honestly self-report drug use when they did engage ^[4] .

In addition, the data linkage component of the recent Older People with Drug Problems in Scotland ( OPDP) project identified a group of people who were using benzodiazepines problematically, but who were not using opioids: women aged 35 or over were more prevalent among this group than among the entire cohort used for the project. A very high percentage of the older women identified in this group did not have any specialist drug treatment or drug-related hospital activity during the study period. Secondly, the National Drug-Related Deaths Database found that the proportion of female decedents who were known to use drugs ^[5] has declined in recent years, from 90% in 2009 to 80% in 2014 (Information Services Division Scotland, 2016a). In contrast, among men this proportion has remained largely stable, with 89% of decedents in 2009 and 87% of decedents in 2014 known to have used drugs. In keeping with this, a small number of informants raised concerns about local examples of drug-related deaths associated with prescription drugs (such as opioid painkillers and benzodiazepines) among older women not fitting the typical profile of people considered to be at risk.

However, against this, other stakeholders had not observed any such trend, and the vast majority of deaths continue to involve illicit opioids. This possibility may therefore warrant further investigation, perhaps as part of a broader project looking at prescription drug use and/or polysubstance use.

3.5 Ageing among women who use drugs

Age is a known risk factor for DRD (McAuley and Millar, 2017). Ageing among the population of women who use drugs might therefore contribute to an increase in drug-related deaths, particularly if this is occurring to a greater extent among women than men.

As described above, estimates of the prevalence of problem drug use among women in Scotland by age group have not been published previously, partly due to the small numbers involved (Information Services Division Scotland, 2014). Other sources can however give a partial indication of the age profile of those affected and those at risk.

Among clients undergoing initial assessment for specialist drug treatment, a definite ageing effect was observed between 2006/2007 and 2015/16 among both genders. For instance, one simple metric is the proportion of such clients aged ≥35 years. In all years this was greater among men than women, but this proportion grew more steeply over this period among women than among men (Figure 2). Similar results are observed in data from the NESI survey of people attending injecting equipment provision outlets in Scotland between 2008-09 and 2015-16, which includes people not in contact with treatment services (see Appendix 4, additional table 2).

Though this analysis is fairly crude, it is corroborated by analysis from the OPDP project. This found that the gender gap in prevalence of problem drug use was narrower at older ages and that the overall population of older people with problem drug use in Scotland is likely to reach its peak number later among women than among men – partly due to higher rates of treatment success and survival (based on all-cause mortality) among the former.

There is therefore good evidence that the average age of women who use drugs in Scotland is increasing, and some indications that this trend might be more marked among women than men.

This may be particularly relevant given that, as described above in section 2.5, there is some evidence to suggest that the effect of ageing on DRD risk may be more pronounced among women than among men.

In keeping with this, NRS estimates of DRD rates per 1,000 problem drug users for the period 2011-2015 are substantially higher among men than women for those aged 15-34, but the reverse is true for those aged over 35. This is illustrated in Figure 3. It is important to note here NRS’s own caveat that these rates are broad indications only, as the estimated numbers of people with problem drug use may be subject to wide confidence intervals.

Figure 2. Trend over time in the proportion of clients undergoing initial assessment in specialist drug treatment services who are 35 years or older, by gender.

Source: Scottish Drugs Misuse Database, ISD Scotland.

Figure 3. Annual average drug-related death rate (2011-2015) per 1,000 people with problem drug use, by age group and gender

Source: National Records for Scotland. Note that numbers of drug-related deaths are based on data from 2011-2015, and estimated numbers of people with problem drug use are based on data from 2012/13. Note NRS caveat that these rates are broad indications only, as estimated numbers of people with problem drug use may be subject to wide confidence intervals.

Looking at trends over time, between 2000 and 2016 the proportion of drug-related deaths who were aged 35 years or more increased in both genders but in almost all years was higher for female decedents than male decedents (Figure 4). That this is the reverse of the pattern seen for people accessing harm reduction or treatment services (Figure 2) may suggest either that older women are not accessing services or that they are at higher risk of death (or plausibly, and possibly relatedly, both).

Figure 4. Percentage of cases of drug-related death who were aged 35 years or more, by gender

Source: National Records of Scotland

Together these findings suggest an ageing cohort of women who may be increasingly vulnerable to drug-related death, with levels of risk approaching – or exceeding – that of their male peers.

However, a number of factors argue against an ageing effect alone being a sufficient explanation of the increase in drug-related deaths among women. Firstly, DRD rates per 1,000 population have increased or remained largely stable within all age strata over the last 15 years, except for men aged 15 to 24, among whom they have decreased (Figure 5). As described above, data on trends over time in DRD rates per 1,000 people with problem drug use are not available; however, these general population rates suggest static or increasing risk even when age is held constant.

Secondly, it is important to examine why age is associated with an increase in the risk of drug-related deaths, and why this effect may vary by gender. It is unclear to what extent age exerts a direct causal effect on DRD risk (for example, through a greater burden of physical co-morbidities) or is a marker for other risk factors (such as poly-substance use, social isolation, or more complex life circumstances).

Figure 5. Rates of drug-related death per 1,000 population, by gender and age group. (Note different scales of Y axis).

Source: National Records of Scotland

Both stakeholder input and the published literature identified age as a factor that needed to be critically interrogated, arguing that it was important to understand the activities and experiences of those within this cohort (King et al., 2013, Scottish Drugs Forum, 2017, Gao et al., 2016). For instance, stakeholders posited that for people who use drugs, increasing age is associated with increasing isolation, as social networks are depleted through bereavement and other changes in friendship networks and family relationships. This was also a recurring theme in the qualitative work carried out as part of the Older People with Drug Problems project (Scottish Drugs Forum, 2017a), as illustrated by the quotations below. Some professional stakeholders interviewed felt that for some women in particular, mid-life and the menopause may be a particularly challenging life stage as they confront their loss of fertility in the context of limited or no relationship with their existing children, resulting in hopelessness for the future. Relationships, including as parents, are explored further in section 3.7.

"I think it's got quite a lot to do with mental health, but sometimes I don't know, I just, I've given up just now, because everybody's been dying round about me, just giving up."

(Interview 303)

"…now that I'm 35 I'm thinking "Oh my God, I'm nearly hitting 40, I'm still using gear, I haven't got a job, a lot of mental illness, I haven't got any kids, I'm not married. My Mum and Dad wanted more for me than that and I feel, they make me feel guilty about that you know."

(Interview 503)

Ageing effects (which are the consequence of growing older) can also be difficult to separate from cohort effects (which are the consequence of being born during a specific period) and period effects (which are the consequence of factors that occur at a particular time and affect all age groups equally). A recent analysis of time trends in drug-related death risk in Scotland, which tried to disentangle these factors, found that as well as ageing, a cohort effect associated with an increase in risk among those born between 1960 and 1980, was present among women but less pronounced than among men (Minton, 2017).

Age may therefore be important not only as a cause of physical vulnerability but also as a marker of wider life changes and transitions as well as historical experiences. Many of these age-related factors – whether physical, mental, or social – are potentially amenable to intervention and support. Some are explored further in subsequent sections

3.6 Changing patterns of drug use among women

One potential explanation for changes in drug-related death rates might be changes in the type, combination, quantity, composition, frequency, and route of drugs consumed, which may in turn be explained by trends in availability, price, enforcement, or other factors.

Among people presenting for initial assessment in specialist drug services, trends over time in the main illicit drug of use (generally defined as the substance for which the individual is seeking treatment) have been relatively similar between the genders (Figure 6). Opioids remain by far the most common drug for which people seek help from specialist drug services, with presentations increasing recently following a longer-term decline in this drug type between 2006/07 and 2012/13. Looking at the proportion of people by drug type, opioids showed a lesser decline during the period 2007/08-2011/12 among women than men, and remained the main illicit drug among a higher proportion of women than men in recent years (e.g. 63% vs 54% in 2015/16). Both sexes showed slight increases over the long-term in the number of individuals seeking help for problems with sedatives (primarily benzodiazepines) and cannabis. Both sexes showed increases in the number of individuals seeking help for problems with "other drugs", though this was more marked among men. These data are limited to a single drug, so will not capture instances where individuals are concurrently using multiple drugs. They are also based on individuals seeking treatment, though similar results were found in the NESI survey of people attending injecting equipment provision outlets in Scotland, which is likely to include a broader range of people than SDMD data.

Throughout the period 2008-09 to 2015-16, the NESI survey found that a similarly high proportion of both women and men reported having injected heroin in the last six months (97% women compared to 94% men over that period), though women’s use of heroin showed a less pronounced dip in the years 2011-12 and 2013-14. Women taking part in NESI were less likely than men to inject cocaine, crack, or heroin/cocaine in combination, though both sexes showed a similar upward trend in cocaine and crack use in recent years. For injected speed (amphetamines) and injected benzodiazepines, numbers were small and there were no clear trends by gender. Data on legal high injecting have only been collected in one sweep of the survey to date but no gender differences in the prevalence of use were identified (10% among both men and women). Women participating in NESI were somewhat more likely to have received prescribed methadone in the last six months than men (82% vs 75% over the period 2008-09 to 2015-16), with fluctuating trends over time in both genders. This gender difference was also observed when restricting the analysis just to people currently injecting.

Figure 6. Number of individuals presenting for initial assessment in specialist drug services by gender, main type of illicit drug, and financial year. (Note different scales of Y axis).

Source: Scottish Drug Misuse Database, ISD Scotland

The following paragraphs describe findings from death registration data: for the purposes of this analysis, data on deaths meeting the Office for National Statistics ( ONS) ‘wide definition’ of drug-related deaths were also reviewed, alongside the more limited number of deaths meeting the NRS definition, in order to examine the potential role of a wider range of drugs (including those currently or previously uncontrolled). When interpreting these data, it is important to bear in mind that determination of drug implication in deaths is complex and requires an element of subjective judgement on the part of the toxicologist, particularly for newer drugs. This is discussed further in a recent review of the role of benzodiazepines in drug-related mortality, undertaken by NHS Health Scotland (Johnson et al., 2016).

Death registration data shows that opiates are still the most commonly implicated drug in both sexes, and that they account for a substantial proportion of the increase in recent years (Figure 7).

When this is broken down by opiate, heroin/morphine and methadone are the most commonly implicated drugs (whether alone or in combination). The number of deaths in which heroin/morphine is implicated have increased substantially since the early part of this decade, in both sexes. This may reflect the effect of the heroin shortage in 2010/11, but in both sexes the number of deaths in which heroin/morphine is implicated have now risen above pre-shortage levels.

Among women, the proportion of deaths involving methadone has in most years been slightly higher than among men, and the proportion of deaths involving heroin/morphine slightly lower. This is consistent with the somewhat higher proportion of actively injecting women who are receiving prescribed opioid substitution therapy ( OST) observed in the NESI study (79% compared to 72% between 2008-09 and 2015-16; Health Protection Scotland, 2017a) and the higher proportion of female drug-related deaths who were prescribed OST at the time of death (Information Services Division Scotland, 2016a). Alternative (or concurrent) explanations might include higher rates of illicit methadone use or a greater physical susceptibility to complications of methadone treatment.

With regard to trends, there has been a gradual upward trend in the number and proportion of deaths in which methadone is implicated in both sexes, with an apparent jump in the last one or two years. The National Drug-Related Deaths Database contains information on methadone dosage, duration of treatment, and on the proportion of deaths where methadone was present who had an active substitute prescription: an analysis of gender differences in these factors might be a fruitful area for further analysis.

With regard to other opiates, deaths implicating codeine and dihydrocodeine have shown a slight increase in recent years among women decedents whilst remaining largely stable among men.

Other trends appear to be common to both genders. Since around 2014, there has been a pronounced rise in the number of deaths in which benzodiazepines are implicated among both men and women, and a lesser increase in the number of deaths in which cocaine is implicated (Figure 7).

The latest NRS data available indicate that novel psychoactive substances ( NPS) implicated in drug-related deaths are almost exclusively benzodiazepine-type drugs, such as etizolam, and are almost always found in combination with other drugs (most commonly heroin and methadone). However, some stakeholders identified relatively recent concerns about increasing use of NPS among people with existing drug problems, which they linked to a growth in underground markets following the introduction of the Psychoactive Substances Act 2016. These trends may not yet be detectable in published drug-related deaths registrations but may be an area for ongoing vigilance.

Figure 7. Number of deaths where opiates, benzodiazepines, cocaine, or alcohol were implicated in, or potentially contributed to, the cause of death between 2000 and 2016, by gender: NRS definition*. (Note different scales of Y axis).

Source: National Records of Scotland

* NRS definition of drug-related death, based on UK Drug Strategy. These data will therefore not include deaths involving any substances uncontrolled at the time of death (e.g. an overdose of tramadol alone prior to 10 June 2014 or an overdose of etizolam prior to 31 May 2017). The dashed line delineates a change in reporting practice for drugs involved: up to 2007, some pathologists reported only those drugs which they thought directly contributed to the death, whereas from 2008, they report separately drugs which were implicated in, or which potentially contributed to the death (shown here), and those which were present, but were not considered to have contributed to the death (not shown here). Since these data record individual mentions of particular drugs, there will be multiple-counting of deaths where more than one drug is present.

Prescription drugs were identified by a number of stakeholders as a potential factor in the observed trend in DRDs among women, in two respects.

The first is the problem use of prescription drugs by women without a history of dependence on other drugs and who do not fit the conventional profile of people at risk of drug-related harms. This was discussed earlier, in section 2.1.

The second is prescription drug use (whether prescribed to that individual or obtained through illicit supply) among women with established problem use of street drugs who are typically known to addiction services and may have other complex needs, such as homelessness or trauma.

Prescription drug use was mentioned by a number of stakeholders, particularly in the context of polysubstance use and interactions with opiate substitution therapy. Drugs mentioned as being of particular concern in recent years and months include pregabalin and gabapentin and benzodiazepines, especially short-acting formulations such as alprazolam (Xanax). Some stakeholders felt that women were more likely to be prescribed drugs with the potential for abuse such as gabapentinoids or opiate pain-killers, though illicit supply was also recognised as an important source. One stakeholder identified that prescription drugs – which are not detected on urine screening – may be particularly appealing to women, given the potential implications of a positive urine screen for child protection proceedings or custody arrangements.

Using the ONS ‘wide’ definition for drug-related deaths (which includes drugs not controlled at the time of death), there appear to be similarities and differences between the genders in deaths involving prescription drugs. As described above, there has been a pronounced increase in the proportion of deaths involving benzodiazepines in recent years, in both sexes. Deaths involving gabapentin have also shown a marked rise in both sexes, though this trend appears to have started slightly earlier among women than men (around 2010 vs 2012) and the percentage of deaths in which gabapentin is involved has consistently been higher among women than men. Both sexes have also seen slight gradual upward trends in the number of deaths with anti-depressants and/or anti-psychotics involved. These trends are illustrated in Figure 8. (Note that these figures are based on the ONS ‘wide’ definition of drug-related death and will therefore include some deaths not counted as part of the NRS definition used elsewhere in this report, and in particular, not counted in Figure 7).

With regard to polysubstance use, NRS death registration data show that the proportion of deaths in which only one drug (and, perhaps, alcohol) was implicated has declined among both sexes since 2007, with this proportion generally lower among women. For instance, in 2016, only 16% of deaths among women involved only one drug (and, perhaps, alcohol) compared to 24% among men.

Toxicology data from the National Drug-Related Deaths Database for the period 2009-2014 shows that the most common combinations of drugs found at post-mortem among women and men were those involving either heroin or methadone with benzodiazepines (Information Services Division Scotland, 2018b). In general, combinations involving heroin were more common among male decedents, whereas combinations involving methadone were more common among female decedents. For most combinations, the direction of the trend over time was similar for both genders, although the increase in heroin with gabapentin or pregabalin and in anti-depressants with dihydrocodeine was more marked among female decedents.

Polysubstance use (whether of illicit or prescribed drugs, or – most commonly – a mix of both) was felt by informants to be particularly problematic with increasing age, due to a loss of physiological reserve and greater prevalence of co-morbidities. It was queried whether biological differences in tolerance might also contribute to a particular vulnerability to polysubstance use among women who use drugs. If women are more likely to be prescribed or otherwise use high-risk prescription drugs, this might interact with the more pronounced ageing phenomenon among women who use drugs described in section 3.5 to contribute to the trend observed.

Informants also acknowledged that patterns of substance use were complex and dynamic, often in ways that may not be captured through routine data: for instance, greater use of cocaine and crack on days when benefits were received.

It is unclear whether gender differences in harms might reflect differences in the prevalence of polydrug use or in vulnerability to its effects. Further work to investigate polysubstance use from routine data sources – such as SDMD and death registrations – is likely to be a valuable area for future work.

Although one stakeholder interview identified alcohol as a potential factor in drug-related deaths among women, NRS data on decedents indicates that the number of deaths in which alcohol is implicated has remained largely stable in both sexes over the past fifteen years (Figure 7).

Figure 8. Number of drug-related deaths involving selected prescription drugs between 2000 and 2016, by gender: ONS ‘wide’ definition*. (Note different scales of Y axis).

Source: National Records of Scotland

*The ONS ‘wide’ definition includes all deaths coded to accidental poisoning, and to intentional self-poisoning by drugs, medicaments and biological substances, whether or not a drug listed under the Misuse of Drugs Act was present in the body. The dashed line delineates a change in reporting practice for drugs involved: up to 2007, some pathologists reported only those drugs which they thought directly contributed to the death, whereas from 2008, they report separately drugs which were implicated in, or which potentially contributed to the death (shown here), and those which were present, but were not considered to have contributed to the death (not shown here). More than one drug may be reported per death. These are mentions of each drug, so do not add up to the overall total number of deaths.

3.7 Changes in relationships and parenting roles

This section investigates potential explanations relating to relationships, whether parent-child, family, social, or intimate.

The link between loss of child custody and risk of drug-related death for women was a recurring theme in conversations with professional stakeholders and in the interviews with women who use drugs undertaken as part of the OPDP project. Both sources suggested that loss of child custody was closely linked to loss of motivation for recovery, feelings of hopelessness, and increased risk of drug-related death: this is supported by the limited literature available on this topic (as described in section 2), and by work in progress on this topic by researchers at the University of Glasgow (Russell, 2018).

With increasing age, women may have experienced multiple child removals, may have very limited positive family relationships, and may find the prospect of declining future fertility in this context difficult. Some professional informants felt these factors were contributing to feelings of hopelessness and risky consumption behaviours among women, which increased the risk of drug-related death. On the other hand, one informant suggested that without the risk of pregnancy and fear of child protection concerns, older women may be more able to engage with treatment services.

One stakeholder suggested that reorganisations in treatment services locally in their area may have increased the number of child removals, and therefore the risk of death among women, due to lower ‘thresholds’ around child protection issues in the new teams. However, it should be acknowledged that this was only one informant’s view: other informants did not feel that child removals were increasing, but recognised it as an ongoing concern.

With regard to routine data on this topic, data from the National Drug Related Deaths Database show that the proportion of those dying from drug-related causes between 2009 and 2014 who were a parent or parental figure was somewhat higher among women than men (44.3% vs 34.1%) (Information Services Division Scotland, 2018b). Women dying from drug-related deaths were also more likely to live with children at the time of death than men (13.5% vs 6.1%). The proportion of women who died who were parents or parental figures, and the proportion living with children, increased between 2009 and 2011 and subsequently declined; among men who died, both measures have been relatively stable over time (Figure 8).

Figure 8a. Percentage of drug-related deaths where the deceased was a parent or parental figure to children under 16 years, by gender and year.

Source: National Database on Drug-Related Deaths, ISD Scotland

Figure 8b. Percentage of drug-related deaths where the deceased was living with children under 16 years at the time of death, by gender and year.

Source: National Database on Drug-Related Deaths, ISD Scotland

Scottish Government social work statistics show that the number of children ‘starting to be looked after’ (i.e. taken into local authority care, in whatever setting, for the first time) increased slightly between 2003 and 2009, before subsequently declining to reach a fifteen-year low in 2017 (Scottish Government, 2018). Although data are not routinely published on trends in the nature of the child protection concern, figures from 2016 indicate that parental substance misuse (of alcohol and/or drugs) is a recorded concern in 39% of case conferences for children on the register.

However, these data – relating only to parental status of cases of drug-related death or aggregate population statistics on child protection – are limited in what they can reveal about the lived experience of parenting roles and social work involvement, and about trends over time. This is therefore an area that may merit further investigation, as described in section 4.3.

With regard to relationships more generally, increasing social isolation was identified as a growing problem by a number of stakeholders, and has also been highlighted in the OPDP qualitative work (Scottish Drugs Forum, 2017a). This was felt to be explained by deaths within peer groups, relationship breakdown with non-using friends and family among people who continue to use, the end of friendships with drug-using peers among people entering recovery, mental health problems, or a lack of trust in others following previous traumatic experiences. Some of these themes are illustrated by the quotations below.

These factors may be particularly salient for women: as described in section 2, women’s drug use tends to be more closely linked to intimate relationships and is more highly stigmatised. One informant suggested that women who use drugs experience greater downward ‘social drift’ than men and therefore a greater loss of social bonds over the course of their drug use, though a brief literature search did not find any evidence to confirm or refute this suggestion. Some stakeholders also highlighted a lack of gender-concordant peer support for women within existing services. On the other hand, there is some evidence to suggest that women are more able to establish new social relationships unrelated to drug use than men, and enjoy a greater level of practical and emotional support from family members (Neale, 2004), suggesting that social isolation may be less of a problem for women than men.

“. … move back to [place], so I can see my mum, and my family and my daughter and my grandchildren, and, because they all just think that I’m, you know, really f***** up, I am, you know what I mean, I am, I don’t lie to them, but I don’t know, it seems to be better when I’m around them, me and my mum are so alike, I couldn’t stay with my mum, because I stayed there for a wee while, but in the end, we end up fighting, and it’s the same with my sister, my dad, he’s only like got one room, so that’s no good, and he’s not talking to me at the minute, because of my sister, so I feel like I’ve just got nobody really”

(Interviewee 318)

“So I needed [name], because my mum wasn’t well, because obviously I’ve cut everybody out of my life that I did, anybody I knew that was all to do with drugs.”

(Interviewee 126)

3.8 Experience of adversity, trauma, and violence

Although experiences of adversity, trauma, and violence are relatively common among women who use drugs (Section 2), and may increase the risk of drug-related harms, identifying whether they might contribute to trends over time in mortality rates is more challenging. For instance, there is very little longitudinal evidence available from the published literature with which to understand trends.

The only source of routine data available on this topic is the National Drugs-Related Deaths Database, which collects information on whether the decedent was known to have been a victim of domestic violence or sexual abuse. Among those dying between 2009 and 2014, both of these experiences were much more common among women than men (74% vs 6% and 50% vs 8%, respectively), but there were no clear trends over time observed in either gender (Information Services Division Scotland, 2016b).

Stakeholders, however, did identify some potential factors which have changed in recent years which may interact with women’s experiences of adversity, trauma, and violence to increase the risk of drug-related harms, including death. These are noted below, though many are explored in greater detail in subsequent sections.

• Changes in the welfare benefits system which make women more vulnerable within relationships. For instance, sanctions may increase women’s financial dependence on partners, or force them to engage in commercial sex work. Another example cited was the move within the new Universal Credit arrangements for payments to be made to a single ‘head of the household’ rather than individual household members, though it should be noted that this is a relatively new change and not yet fully rolled out across the country.
• Life events or transitions which are more common among an ageing cohort, such as bereavements, multiple child removals, loss of fertility (Section 3.7), which may not only be traumatic in themselves but also compound the effects of previous trauma.
• Changes to health and social services which result in reduced provision (e.g. mental health), a lack of continuity (whether in services themselves or in staffing), or a less holistic approach (e.g. as a result of reduced staffing and skills within addictions services or more ‘punitive’ approaches). These changes particularly may affect engagement and outcomes among people with a history of trauma, who can experience particular difficulties in accessing services or establishing therapeutic relationships.

3.9 Co-occurring physical and mental health conditions

There is evidence that diagnosed mental health problems, prescribed psychotropic medications, and feelings of indifference and carelessness are risk factors for drug-related death (European Monitoring Centre for Drugs and Drug Addiction, 2012). There is also evidence that some physical conditions – such as hepatitis and cirrhosis – are also associated with increased risk (European Monitoring Centre for Drugs and Drug Addiction, 2012).

As described above, the link between age and concurrent physical and mental health conditions may be one of the mechanisms by which an ageing cohort of people who use drugs is associated with an increase in DRDs (Scottish Drugs Forum, 2017). A high proportion of interviewees from the OPDP project described co-existing mental and physical health problems, many of which were untreated or under-treated. Some of the barriers to treatment have already been described above in previous sections, including stigma, concerns about confidentiality, and difficulties attending scheduled appointments (and potentially being removed from practice lists as a result). Other factors mentioned were a perceived lack of time among GPs for patients who use drugs and poor continuity of substitute prescribing during periods of inpatient care.

“My mental health should be dealt wi’ more…eh the stigma. Doctors should listen to you more and… I think mainly older drug users, our bodies are all physically messed up wi’ arthritis and blood clots and whatever, we should have, right, we should have mair [more] checks on our bodies a couple o’ times a year to get checked out. A full MOT and all.”

(Interviewee 106)

“My partner, he’s saying to me, I think I’ve got that COPD [chronic obstructive pulmonary disease] or something, at night, it’s all bubbling and he said, he says it’s really scary, but, and I know I am, I get out of breath so easy, and it’s just wheezing...”

(Interviewee 421)

"And health ways, when you're older and using heroin, I think you worry more about your lungs with smoking it, then you're worrying about organ failure, if you're injecting it, you know, stuff like that, because you get that the older, you know, you start worrying yourself, you know, about your health, and they say down at [name], that now they find that around about 42 year old, that organ failure can, it can start, you know, about that age, if you've been using long term"

(Interviewee 201)

Interviewees also described the link between life circumstances and physical health (for instance, being underweight because of a dislike for the food in temporary accommodation, or living on benefits) and between mental and physical health (for instance, depression causing weight loss).

During the period 2009-2014, female decedents were more likely than male decedents to be long-term sick or disabled (21% vs 17%), to have at least one physical health condition noted in their healthcare records in the six months prior to death (68% vs 54%) and to have had an acute hospital stay in the six months prior to death (Information Services Division Scotland, 2016a).

With regard to trends, the proportion of women who had experienced a recent physical health condition prior to death rose from 49% in 2009 to 76% in 2014 (Information Services Division Scotland, 2016a). For instance, the proportion of female decedents with a respiratory condition rose from 24% to 36% over the same period, compared to 16% to 20% among male decedents.

Female decedents were also more likely than male decedents to have at least one psychiatric condition noted in their healthcare records in the six months prior to death (66% vs 47%), and had a higher average number of psychiatric conditions. The mean number of diagnosed psychiatric conditions among decedents increased between 2009 and 2014 to a similar extent in both genders (women; 0.7 in 2009 to 1.2 in 2014, men; 0.5 in 2009 to 0.9 in 2014).

Data from drug treatment services can be used to ascertain to what extent these findings reflect trends in the population at risk. Among individuals attending drug services in Scotland for initial assessment or follow-up ( SMR25a or SMR25b), the proportion who report co-occurring health issues has increased slightly among both sexes since 2006/07: this increase has been more pronounced among women than among men (Information Services Division Scotland, 2018a). For instance, 46% of both male and female clients in 2006/07 reported a co-occurring health issue: by 2015/16, this had risen to 54% of male clients and 62% of female clients.

When these figures are broken down by specific health problems, the proportion reporting drug-related physical health problems was slightly but consistently higher among female compared to male clients, with both sexes showing a small increase in recent years (to 32% among women and 30% among men). The proportion reporting mental health issues has increased among both sexes, though this trend was more pronounced among women (22% in 2006/07 to 37% in 2015/16, compared to 20% to 27% among men). There was no clear trend in the proportion of clients reporting alcohol problems: rates continued to be higher among men but the data suggest a slight convergence in recent years.

These data are corroborated to some extent by a linkage exercise carried out as part of the OPDP project, which examined hospital admissions among a cohort of older people with drug problems identified from SDMD data, prescribing records, and drug-related hospital episodes (Scottish Drugs Forum, 2017). This found that admissions for asthma/chronic obstructive pulmonary disease ( COPD), lung cancer, and depression were much higher among older women with drug problems compared to their male peers, and compared to women in the rest of the population. In contrast, older men with drug problems tended to have higher rates of hepatitis C, liver disease, psychosis and heart disease. These data suggest that co-morbidity varies with gender and with age, and point to some factors that may contribute to differential risk of drug-related death (for instance, poorer lung health among women).

More generally, poor mental wellbeing and sense of hopelessness among women who use drugs was mentioned as an important factor by a number of stakeholders. This hopelessness was felt to increase the risk of drug-related death, either directly by increasing the risk of deliberate (or ‘ambivalent’) overdose, or indirectly, by precipitating increased or more risky consumption patterns without any particular intent at self-harm. Several precipitating factors were described, including loss of access to children, bereavement, and changes to welfare benefits. Women interviewed as part of the OPDP project also identified a reciprocal relationship between social isolation and mental health.

There is wider evidence from population-based surveys that mental health has worsened in recent years among people living in the most deprived areas of Scotland and among people receiving welfare benefits such as Income Support/Job Seeker’s Allowance (Scottish Public Health Observatory, 2017c) – both of which are likely to apply to a significant proportion of people who use drugs.

One participant in the Older People with Drug Problems project described how challenging ‘everyday’ tasks such as sorting out housing and benefits could be, and their effect on her mental health:

“I get panic attacks. Sometimes it’s a bit you know, I don’t really know why. They just come out of the blue you know. If I’ve got too many things to do, you know like say they go, “Right you’ve got to do this this week, and that this week” and I just panic cos I think somehow it’s, I don’t know why, cos, once I do it I can actually get it done. But I just go into this panic mode, changing address, ah, and it’s a big, and then once I’ve done it I think, “Why were you getting in a panic over that? It was only a couple of phone calls”. But I suppose because your life is so outside everything like that when you’re using all the time. When you suddenly have to do these things, you get panicky about them.”

(Interviewee 213)

During 2009-2014, suicides accounted for a greater proportion of drug-related deaths among women than men (14% vs 5%) (Information Services Division Scotland, 2016a). However, despite indications of an increase in mental health problems among this population, the increase in drug-related deaths among women in Scotland does not appear to be driven by an increase in suicides (see Appendix 4, additional table 2). The proportion of DRDs attributed to intentional self-poisoning among women has fluctuated over the years but the general trend is a stable or declining one, reflecting a small increase in absolute numbers that is outstripped by a much greater rise in deaths attributable to accidental poisoning and drug abuse. However, intentionality can be hard to define in this context, with key informants reporting that ambivalence of intention was common among people at risk of, or who have survived, overdose.

These data should be interpreted with a number of caveats: for instance, women in the general population tend to have a somewhat higher prevalence of limiting long-term illness and self-reported mental ill-health compared to men (Scottish Public Health Observatory, 2018, Bardsley et al., 2017), and there are limited data from sources other than treatment services or death registrations. However, together they do suggest that a greater increase in the prevalence of concurrent physical and mental health conditions among women who use drugs than their male peers, which may be contributing to increases in drug-related deaths.

3.10 Access to and engagement with treatment and harm reduction services

Sustained engagement with treatment services, including opiate substitution therapy ( OST), is an important protective factor against drug-related death (European Monitoring Centre for Drugs and Drug Addiction, 2012, Mathers et al., 2013, Dickie et al., 2017). Conversely, there is good evidence that transitions in and out of treatment can increase risk of death among people who use drugs (European Monitoring Centre for Drugs and Drug Addiction, 2012, Dickie et al, 2017).

As described in Section 2, there is mixed evidence about women’s access to and engagement with treatment relative to men, particularly from the UK.

Both the total number and proportion of clients undergoing initial assessment for specialist drug treatment who are female have remained relatively stable over the past ten years in Scotland (Information Services Division Scotland, 2017d). The proportion of women among people undergoing initial assessment in drug services over this period (28-31%) has been similar or slightly higher than the proportion of women observed in the NESI survey samples between 2008-09 and 2015-16 (ranging from 27-30%) and in the most recent estimates of the prevalence of problem drug use in Scotland (30%) (Health Protection Scotland, 2017b, Information Services Division Scotland, 2014). The fact that women are equally, or better, represented in data sources based on treatment services compared to data sources based on community surveys could be interpreted as an indication that, in Scotland in recent years, women with drug problems are equally or more likely to access treatment services than men.

There is some further evidence from the National Drug-Related Deaths Database to support this. Between 2009-2014, a slightly greater proportion of female decedents had been in contact with drug treatment services in the six months prior to death than male decedents (50% vs 42%) and to have been prescribed opioid substitution therapy at the time of death (38% vs 23%) (Information Services Division Scotland, 2016a). Female decedents were also more likely than male decedents to have experienced previous overdose (58% vs 48% for at least one, and 26% vs 15% for five or more).

These data may support the idea that for women, the problem may lie more with services not meeting the needs of clients or missed opportunities to intervene, than with difficulties in access. This was echoed by some stakeholders. It is also consistent with the observation from NESI that women attending injecting equipment provision outlets were more likely to have received prescribed methadone in the last six months than men, indicating ongoing street drug use despite opioid substitution therapy.

The interviews with older women with drug problems identified a number of factors associated with women leaving or being dissatisfied with treatment, including perceived under-dosing, reluctance to reduce dose, being on OST too long, and lack of support: some of these themes are illustrated in the questions below. Lack of capacity and time at drug treatment services was also mentioned. Conversely, some interviewees described positive experiences with services and support groups, though lack of awareness or misunderstandings about eligibility were cited as a barrier.

With regard to harm reduction interventions for drug-related death, the main source of information is naloxone supply and utilisation. Trends between 2008-09 and 2015-16 in the proportion of NESI survey participants reporting having been prescribed take-home naloxone ( THN) were very similar by gender, with both men and women showing a substantial increase in THN receipt in the three successive surveys since 2011-12, from around 8% to 32%. There was a slight trend towards a higher proportion of women having been prescribed naloxone in recent years (for example, 34% among women compared to 31% among men in 2015-16) (Health Protection Scotland, 2017a). However, rates of carriage were low in both genders, and appeared to decline over time (for example, 8% in women and 5% in men in 2015-16). As described further in section 3.11, naloxone uptake among female prisoners on release appears to be relatively high.

"I think sometimes when you've been on methadone for a wee while, and you fall back into heroin, they could put you up, you know, but they're keeping me at 20ml, they're not budging with my 20ml, there's a lot of folk on and then they're set up for 30ml, I'm only 20ml at [name], and they'll no budge and put my methadone up, even though I went back onto heroin for a year, there's still expecting me to get off the heroin, giving me 20ml of meth, which is impossible, you know, I just can't do it."

(Interviewee 201)

"Every time I asked them to drop me down, they wouldn't drop me down, they kept telling me it was too early, but I knew I was ready to get dropped down, you know."

(Interviewee 409)

"They say they're too busy down there, because it is bursting at the seams at [name], because a lot of younger ones are coming in now, you know, and they're more wanting, I think they're more wanting to pinpoint them than the older ones."

(Interviewee 201)

"Right, I found I've got an awful lot of positive help now, which I had no idea was available to me…I never attended any of those groups, even although I knew they were out there, the reason being I never went to them was because I was using, I thought you had to be totally clean before you could use them"

(Interviewee 128)

"Any time you need anybody to talk to, there's always somebody free, they do a drop-in twice a week, they'll help you with benefits, housing, basically anything they can help you with, if you come and ask, they'll open that door to anybody and help them."

(Interviewee 512)

The National Drug-Related Deaths Database found that between 2009-2014, women who died were more likely than men have been in the same room as another individual at the time of death (33% vs 23%, among those for whom data were available), though rates of naloxone availability and administration were similar (Information Services Division Scotland, 2016a). This suggests missed opportunities for resuscitation among fatal overdoses of both sexes.

Although this topic has not yet been explored in Scotland in detail, recent reports by Public Health England and the Advisory Council on the Misuse of Drugs have suggested that changes in drug treatment services in England may have been a contributory factor to rising DRD rates, for instance through disruption to the continuity of relationships with service users and partnership working with other agencies (Public Health England, 2017, Advisory Council on the Misuse of Drugs, 2016).

Though the commissioning and service provision landscape in England is distinct to that in Scotland, stakeholders did highlight the potential impact of recent cuts to local drug treatment services. Examples were also cited of gender-sensitive or specific services which had been discontinued, such as a residential family unit in Glasgow and a women’s homelessness service in Edinburgh. Lack of funding was felt to create high thresholds for support that meant that services could only deal with individuals in crisis, rather than on a more proactive basis, and to favour services with lower paid and therefore less skilled or experienced staff. More generally, short funding periods and the commissioning cycle were reported by informants to be challenging in services that depended on relationships, continuity, and trust.

Some stakeholders also raised the role of recovery-oriented approaches to policy and practice. The Public Health England report on drug-related deaths noted that, although no direct relationship could be established, poor recovery-oriented practice may put people at greater risk (Public Health England, 2017). A number of staff from services and academia described gaps in harm reduction provision and potential unintended consequences of recovery-oriented approaches, such as a drive to reduce or cease OST, or a drugs-focused approach less able to address holistic issues around social inclusion, housing, employment, and mental health. Some felt that women might be more vulnerable to the potential risks of a recovery-oriented treatment system, due to the greater salience of stigma and child protection concerns, and relational factors influenced by previous trauma.

Some stakeholders also felt that women were not recognised or accommodated within the growing recovery communities and networks across Scotland, though others identified examples of women-specific groups. However, it was generally agreed that women (and men) may be vulnerable within recovery settings to exploitation, coercive or abusive relationships, or stalking, especially given that recovery processes often involve discussing personal experiences and information. Some of these networks were felt to be under-resourced and perhaps to lack appropriate governance.

Concerns were also raised by some stakeholders about recent reductions in provision of allied services which support people with drug problems, such as education, training, and employability support. Lack of access to these, either due to service cuts or funding restrictions (e.g. lack of funds to cover travel expenses), were felt to hinder women’s efforts to achieve sustained reductions or cessation in drug use.

As people who use drugs age, their health and social circumstances are likely to become more complex (Scottish Drugs Forum, 2017), and they are likely to come into contact with a broader range of services than just drug treatment. The complex health and social needs of this cohort – and the ‘treatment burden’ involved in managing those needs – may affect their engagement with addictions care, as well as vice versa. The Public Health England report noted that for many cases of drug-related death, their contact with services was characterised by frequent movement between drug treatment, mental health care, physical health, and housing support. This was also evident from the National Drug-Related Deaths Database for Scotland and was particularly marked for women, who were more likely than men to have had contact with services other than drug treatment, such as social services and mental health care, in the period leading up to death. This is likely to be an ongoing or increasing issue given the increasing prevalence of concurrent physical and mental health problems among this cohort (section 3.9) and their vulnerability to wider economic trends and welfare reform initiatives (section 3.12). Such interactions and transitions may again represent opportunities for intervention and for greater co-ordination of service provision.

3.11 Experiences of prison and liberation

Recent release from prison is one of the most potent risk factors for drug-related death (European Monitoring Centre for Drugs and Drug Addiction, 2012). Imprisonment was generally recognised by stakeholders as a difficult and disruptive event for women who use drugs, which was often accompanied by the loss of child custody, relationship breakdown, housing difficulties, and deteriorations in mental health.

In 2013, 53% of female prisoners surveyed in Scotland reported drug taking being a problem on the outside; this had decreased to 38% in 2015 (Scottish Prisons Service, 2016). The number of women imprisoned increased by 46% between 2003-04 and 2013-14, outstripping the rate of increase observed among the male prison population; the average length of women’s sentences also increased somewhat over a similar period (Prison Reform Trust, 2017). Although one stakeholder queried whether women are more likely to receive custodial sentences for drug-related offences compared to men, no evidence to test this suggestion was identified.

Problems experienced by women in the criminal justice system were mentioned by a number of informants, including cessation of OST on entry to prison and lack of throughcare support for issues like drug treatment, housing, and relationships.

However, there is some evidence that women’s prisons are performing relatively well on other aspects of harm reduction: uptake of take-home naloxone at the point of release appears to be significantly higher among female prisoners than male. Women account for approximately 5% of the average daily sentenced prison population in Scotland but between 18% to 32% of the naloxone kits supplied in this setting over the last five years (Information Services Division Scotland, 2017c). However, whether this observation has an impact on drug-related deaths will depend on a multitude of other factors, including the likelihood of kits being carried and appropriately used, and the likelihood of overdoses occurring in the presence of others: it may be that higher supply among women is primarily likely to benefit their partners. Gender differences in the availability and use of naloxone among people experiencing drug-related deaths are discussed in section 3.10.

The latest available data from the National Drug-Related Deaths Database shows that, during the six-year period 2009-2014, female decedents were less likely than their male counterparts to have been in prison in the last six months (8% vs 17%) or ever in their lives (32% vs 56%) (Information Services Division Scotland, 2016a). The report indicates that the prevalence of prior imprisonment among people experiencing a drug-related death fluctuated over this period for both genders, suggesting no clear trend.

3.12 Economic and social trends, including austerity and welfare reform

There is a growing evidence base about the short- and long-term impact of economic trends and policy on health and health inequalities (Parmar et al., 2016, NHS Health Scotland, 2017, Glonti et al., 2015). Some of these impacts appear to differ by gender: for instance, national economic crises appear to have a greater negative effect on mental health among women than men, though for suicide the reverse is true (Glonti et al., 2015) .

The health impacts of economic change may be mitigated or exacerbated by policy responses, including public spending on healthcare, social security, and other public services ( NHS Health Scotland, 2016, Scottish Public Health Network, 2013). In the UK, the 2008 recession has been followed by a programme of public sector austerity and changes to the welfare benefits system.

Several reports have highlighted the gendered impacts of these policies, with women disproportionately likely to be affected by cuts to public services and changes in the availability, generosity, conditionality, and administration of welfare benefits (Scottish Government, 2013, Engender, 2015, Rubery, 2015).

People who use drugs may also be particularly vulnerable to these changes, for a number of reasons. Firstly, they are more likely to live in the most deprived areas of Scotland, which are often disproportionately affected by current and historical economic policies, particularly austerity; secondly; they typically experience extremely precarious financial and social circumstances, with high rates of unemployment and welfare benefit receipt; and thirdly, they often have complex health and social needs requiring care from statutory services.

Women who use drugs therefore experience an intersection of structural factors which may heighten vulnerability to the adverse impacts of welfare reform. Two major reports on drug-related deaths have alluded to the impacts of welfare reform in their conclusions. The recent Advisory Council on the Misuse of Drugs ( ACMD) report on rising opioid-related deaths identified as a potential contributing factor the deepening of socioeconomic deprivation in the UK since the 2008 financial crisis, with its associated implications for individuals, treatment services and the welfare system (Advisory Council on the Misuse of Drugs, 2016). For instance, it noted that in England, recent increases in deaths have been greatest among most deprived areas, which are also those which have had the greatest reductions in local authority funding and working-age adult welfare benefits. Local stakeholders for the Public Health England ( PHE) report on drug-related deaths also highlighted delays or withdrawals in benefits, which not only have a direct impact on individual health and substance use but may be followed by lump sum repayments that exacerbate short-term risk (Public Health England, 2017). Although the latter report focused on trends in England, the reserved nature of welfare benefits during the period of interest means that some of its conclusions may also apply in the Scottish context.

This was echoed by stakeholders, who highlighted how past and present social and economic policies, including the welfare system, have had a significant role in shaping the lives of people who use drugs.

One service manager went as far as to state that poverty was the single biggest explanatory factor in the rise in drug-related deaths in recent years. A number of informants were concerned that changes to the benefits system – such as conditionality, sanctioning, and payments to a household as a whole rather than individual members – may have increased women’s vulnerability to abusive relationships, sexual exploitation, or commercial sex work. Another stakeholder working with families affected by substance use described examples of women being sanctioned for being five minutes late to an appointment after taking a child to the toilet. One informant felt that women are likely to be less assertive in accessing benefits they are entitled to, particularly in a context which is increasingly hostile for claimants. Another informant mentioned austerity-driven cuts to drug treatment and mental health services as a potential factor in rising deaths: the latter might particularly affect women who use drugs given their higher prevalence of common mental health conditions.

Among women interviewed as part of the OPDP project, several had had benefits withheld or cut, largely as a result of missing appointments. In one case poor mental health was explicitly cited as the cause of missed appointments and the resulting cuts to benefits.

It is therefore plausible that austerity and welfare reform may play a role in rising rates of drug-related death, and that women may be particularly vulnerable in this regard.

One specific concern raised was rates of homelessness among people who use drugs, particularly in the context of cuts to housing benefits and welfare reform more generally.

Data from the National Drugs-Related Deaths Database found that between 2012 and 2014, there was a slight increase in the proportion of women dying from drug-related deaths who had been in contact with homeless services in the six months prior to death, from 19% to 23% (Information Services Division Scotland, 2016a). However, between 2009 and 2014 the proportion of women who were homeless or living in temporary accommodation at the time of their death was much lower (ranging between 0.0% and 2.2%), and fluctuated without any clear trend – as did the figure for men (Information Services Division Scotland, 2016b).

Looking at NESI data from people attending injecting equipment provision outlets, the proportion of respondents reporting homelessness in the last six months was generally lower among women than men, but fluctuated among both sexes between 2008-09 and 2015-16, with no clear trend over time (Health Protection Scotland, 2017a).

These data suggest that homelessness is unlikely to be a major factor in the trends observed but it will be valuable to see updated data on this question from the forthcoming National Drugs-Related Death Database report, and additional investigation – for instance, by analysis of housing status among SDMD clients – may be warranted to explore this possibility further.

3.13 Summary of explanations

As described above, the question of what might explain a disproportionate rise in drug-related deaths among women is methodologically challenging, and is likely to have a complex answer, involving multiple interacting factors.

From this scoping, the following factors have been identified as potentially contributing to the trend. Again, these are not presented in order of importance or certainty.

• Ageing among a cohort of women who use drugs, which may be more pronounced among women than men and which may act to increase the risk of drug-related deaths through a range of factors, whether individual (e.g., co-morbidities), clinical (e.g. polypharmacy), and social (e.g. bereavements, loss of maternal role).
• Changes in patterns of substance use, particularly polysubstance use and potential increases in the problem use of specific prescription medications (whether prescribed directly or obtained illicitly). These changes might interact with the ageing phenomenon to further increase risk.
• Increasing prevalence of physical and mental health problems, which appears to be more pronounced among women who use drugs than their male peers; this may be linked to the ageing phenomenon described above.
• Changes in relationships and parenting roles, including social isolation and the potential cumulative impact of multiple child removals.
• Ongoing risk among women engaged with drug treatment services, potentially reflecting failures to meet needs or missed opportunities.
• Changes to treatment services, and wider health and social services, particularly cuts in funding resulting in withdrawal of services, reduced provision, under-staffing or under-skilled staffing, lack of continuity in relationships, or a change in ethos.
• Unintended consequences of the policy and practice focus on recovery-oriented systems of care, which some stakeholders felt may have adversely affected harm reduction provision and efforts to address broader life circumstances.
• Changes in the welfare benefits system which may particularly impact on women, and which may interact with:
  • Vulnerability to abusive or coercive relationships
  • Involvement in commercial sex work
  • Previous or ongoing experiences of trauma
  • Mental health issues
  • Cuts to drug treatment services and other health and social care provision
• Interaction between factors above and known barriers for women engaging with treatment services, which may be practical (such as caring responsibilities) or psychological/relational (such as stigma, coercion, and fear of losing custody of children)
• Interaction between the factors described above and previous experiences of trauma and adversity, which affect vulnerability to new challenges, coping strategies, and access to and engagement with services.

Other factors identified which merit further investigation for their potential role include:

• A possible population of women at risk who may be older, predominantly using prescription drugs, and not otherwise known to services.
• Potential problems with drug treatment and throughcare for women in the criminal justice system.

Factors which seem less likely to explain the trends observed include:

• Changes in reporting or recording practices for drug-related deaths.
• Changes in background mortality risk among women in the general population.

3.14 Limitations

The analysis provided here has a number of limitations, largely due to the time and resource constraints involved in this project:

• A formal literature review was not undertaken: rather, the overview of evidence presented here drew on a bibliographic search for systematic reviews, snowballing from relevant articles, and existing grey literature. A more rigorous and systematic review may be warranted to further investigate some of the areas of interest identified above or the emerging implications discussed in the next section.
• Recruitment of informants took place through existing networks and snowballing, using a mix of purposive and convenience sampling. Although efforts were made to ensure a broad geographical and professional scope, this was a qualitative sample and therefore does not attempt to provide a fully representative approach.
• As described above, routine data sources relating to problem drug use have a number of limitations – particularly for accurately understanding the characteristics and experiences of the population at risk. Only a limited descriptive analysis of existing routine data was undertaken here: some of the areas of interest identified may merit more rigorous and detailed epidemiological investigation.

It is also important to note the potential limitations of the definition of drug-related deaths which is used in published statistics and which has provided the focus for this project. This definition focuses on deaths associated with acute drug use, and will exclude some deaths resulting from indirect or longer-term effects of drug use or other causes of mortality among people who use drugs. Examples might include hepatitis C, HIV, or other infections acquired through injecting, or external causes such as accidents, assaults, and suicides not involving controlled drugs. These other causes of death among people who use drugs may show important differences by gender, which would not be accounted for here, though many of the same themes might apply.