Chapter 10: Individual Person-level Analysis of Drug-related, Alcohol-related and Mental Health Issues
Previous chapters have considered the use of various parts of the health service in isolation – A&E, inpatients, outpatients, prescriptions, mental health services and the Scottish Drugs Misuse services. In these chapters the study found that drugs, alcohol and mental health were recurring themes in the differences observed between cohorts.
In this chapter, we consider the 1.3 million people in the study and look to see who amongst them have any evidence of drug related, alcohol related or mental-health related interactions (hereafter in this chapter simply referred to as drugs, alcohol and mental health) at any point during the period of the study. The following datasets contain information about these issues: prescriptions ( PIS), mental health ( SMR04), Scottish Drugs Misuse Database ( SDMD) and acute admissions to hospital ( SMR01).
For example, someone may have had a prescription for a mental health issue (in the PIS dataset) and also have had an acute admission to hospital in relation to a drug issue (in the SMR01 dataset). When looking at these in isolation, the information available on the PIS record did not indicate that there was also a drug-related admission in the study for that person. In this chapter, no comment is made about whether these health conditions occurred prior or after any episodes of homelessness, or whether the person had these health conditions concurrently.
10.1 Drug, Alcohol and Mental Health related issues amongst the cohorts
In the previous chapters the focus was on the amount of activity in the different cohorts ( i.e. the number of interactions). It was not always clear whether the number of people using each service differed between the cohorts. It was also not discussed whether the people using a particular service are the same people as those who use services analysed in different chapters.
This section explores the proportions of people who have any mental health interactions (prescriptions, acute admissions, or mental-health admissions). The same is then explored for drug-related interactions and alcohol-related interactions. In order to see how these vary between cohorts, the proportions of these are presented (Table 10.1).
Table 10.1: Study individuals with evidence of drug-related, alcohol-related and/or mental health related interactions, by cohort.
|Number of people in each cohort||435,853||435,853||435,853|
|Any mental health||47.9%||25.7%||13.6%|
Almost half of the EHC had mental health interactions, more than the MDC or LDC
48% of the EHC had mental health interactions. This is notably higher than for the MDC (26%) and the LDC (14%). However it shows that for a majority of the people in the EHC there is no evidence of mental-health interactions.
A significant minority of the EHC had drug-related interactions, more than the MDC or LDC
14% of the EHC had drug-related interactions. This is notably higher than for the MDC (3%) and the LDC (0.7%). However it shows that for most of the people in the EHC there is no evidence of drug-related issues.
A significant minority of the EHC had alcohol-related interactions, more than the MDC or LDC
11% of the EHC had alcohol-related interactions. This is notably higher than for the MDC (3%) and the LDC (0.8%). However it shows that for most of the people in the EHC there is no evidence of alcohol-related issues.
More people in the EHC had drug-related interactions than alcohol-related interactions
Around a third  more people in the EHC had drug-related interactions than alcohol-related interactions. This differs from the MDC and LDC, where the proportions are more comparable.
10.2 Overlap of drug-, alcohol-related and mental health issues amongst the cohorts
In the previous section it was seen that a majority of EHC people had no evidence of mental health issues, and the same was true for drug-related and alcohol-related conditions. In this section the overlap between these groups is explored. That is, are the people who have evidence of one of these issues the same people as have evidence of the others. For example it may be the case that the EHC people with evidence of alcohol-related issues are a subset of those who have evidence of drug-related issues, who in turn are a subset of those who have evidence of mental-health issues. This would suggest that a majority of the EHC have no evidence of any of these issues. Conversely it could be the case that these sets are all distinct (no people have evidence of more than one of these). This would suggest that 73% of the EHC would have evidence of (exactly) one of these issues.
In order to see how the people with evidence of these issues overlap, we first examine the proportion of people in each cohort and how drug-related and alcohol-related issues coincide ( Table 10.2).
10.2.1 Overlap between people with drug-related and those with alcohol-related conditions
Here only drugs and alcohol are considered and how they overlap.
Most of the EHC do not have evidence of any of drug or alcohol issues
81% of the EHC do not have evidence of any of drugs or alcohol issues. Thus while it is certainly the case that this proportion is smaller than for the MDC (95%) and for the LDC (99%), it can be seen that most EHC people do not have any evidence of drug-related issues or alcohol-related issues.
More people have both drugs and alcohol than expected were these independent
Given the proportion of the EHC who have evidence of drug-related conditions (14%) and the proportion with evidence of alcohol-related conditions (11%), it would be expected that 1.5% would have evidence of both (if these attributes were independent). However it is observed that 6.1% have both drug and alcohol-related conditions. This shows that these attributes are not independent: people are more likely to have drug-related issues if they have alcohol-related issues, and vice versa.
More of the people who have drugs and/or alcohol have both in the EHC than in the controls
Given that drugs and alcohol are not independent, the group of people who have any drug or alcohol- related conditions is considered (19% of the EHC, 5.1% of the MDC and 1.2% of the LDC). This group is largest among the EHC, as would be expected given the findings in Section 10.1. However it can also be seen that the composition of these groups are also different. Around a third (32%) of these people with either drug- or alcohol-related conditions within the EHC have both conditions, compared with 21% of the MDC and 16% of the LDC.
Table 10.2: Study individuals with drug-, alcohol- or mental health related interactions, by cohort.
|No Interactions with any dataset||51.0%||73.8%||86.2%|
|Mental health only||30.1%||21.2%||12.6%|
|Drugs and mental health||7.6%||1.7%||0.3%|
|Alcohol and mental health||4.4%||1.8%||0.5%|
|Alcohol and drugs||0.1%||0.1%||0.0%|
|Alcohol and drugs and mental Health||5.9%||1.0%||0.2%|
|No alcohol or drugs (no interactions or mental health only)||81.0%||94.9%||98.8%|
|Drugs and/or alcohol (with or without mental health)||19.0%||5.1%||1.2%|
|Drugs and alcohol (with or without mental health)||6.1%||1.1%||0.2%|
|Drugs and alcohol as proportion of drugs and/or alcohol||31.9%||21.4%||16.4%|
|Any of drugs, alcohol or mental health||49.0%||26.2%||13.8%|
|Mental health, and drugs and/or alcohol||17.9%||4.5%||1.0%|
|Mental health, and drugs and/or alcohol as proportion of any of drugs, alcohol or mental health||36.5%||17.1%||7.0%|
|Proportion of drugs/alcohol who also have mental health||94.3%||88.7%||79.8%|
10.2.2 Overlap between people with mental health and those with drugs and/or alcohol
Having considered drugs and alcohol above, this subsection looks at how this overlaps with mental health issues.
A majority of the EHC do not have any of mental health, drug or alcohol issues
51% of the EHC do not have any of mental health, drug- or alcohol-related issues. Thus while it is certainly the case that this proportion is smaller than for the MDC (74%) and for the LDC (86%), it can be seen that a majority of EHC people do not have any evidence of drug-related issues, alcohol-related issues or mental health issues.
Just under a third of the EHC have had a mental health issue which excludes drug- or alcohol-related issues
30% of the EHC have experienced a mental health issue which excludes drug- or alcohol-related issues. This proportion is higher than for the MDC (21%) and for the LDC (13%).
Most of the people with drug- or alcohol-related conditions also have mental health issues
Given the proportion of the EHC who have evidence of any drug- or alcohol-related issues (19% of all EHC), the vast majority also have evidence of mental health issues. In particular most of the people who have evidence of drug- or alcohol-related issues also have evidence of mental health issues, especially in the EHC (94%, compared with 89% in the MDC and 80% in the LDC).
5.9% of the EHC have drug-, alcohol-related and mental health issues
The proportion of people with all three conditions is higher among the EHC (5.9%) than in the MDC (1.0%) or the LDC (0.2%).
In the remainder of this chapter we use a simpler categorization as follows:
- people with no evidence of mental health, drug or alcohol issues (referred to as "None");
- people with all three issues (drugs, alcohol and mental health);
- all other people (these have one or more of drugs, alcohol or mental health issues, but not all three).
10.3 Drugs, alcohol and mental health by age and sex
10.3.1 All three issues - drugs, alcohol and mental health
Figure 10.1 shows the age and sex of people in the cohort who have evidence of all three issues – drugs, alcohol and mental health – at some point during the study period. The age used in this and subsequent figures is the age each person would have attained at 31 March 2015.
In the EHC more of the males than females have all three issues
The proportion of the EHC that has evidence of drugs, alcohol and mental health issues is 7.2% for males and 4.6% for females. This differences is larger than the difference between sexes among the MDC (1.1% males, 0.9% females). The proportion for the LDC are too small to comment upon (<0.2% for both males and females).
In the EHC a higher proportion of males than females have all three issues for ages 26+years
In the EHC a slightly higher proportion of females than males have all three issues for 0–20 years. At 21–25 years the proportions are similar. In each age above this the proportion is higher for the males than for the females ( Figure 10.1).
The proportion of the EHC that have all three issues is lowest for young and old people
At ages 26–60 years (for males, 21–60 for females) the proportion of the EHC that have all three issues is higher than the average across all ages. Above and below this age range the proportion is below the average.
Figure 10.1: Percentage of people in the study with a drug, alcohol and mental health issue at some point during the study period, by age, sex and cohort.
10.3.2 People with no evidence of drug, alcohol or mental health issues
Figure 10.2 shows the proportion in each cohort with none of the above issues at any point during the study period. For the EHC cohort, the proportion of people with none of these issues falls sharply with age. Females fall more sharply than males. By age 36 to 40 years, only around a third of the EHC cohort have none of these issues, and this is much lower than for the MDC and LDC.
Figure 10.2: Percentage of people in the study with no evidence of drug, alcohol or mental health issues during the study period, by age, sex and cohort.
10.4 Repeat and Once-Only Homelessness
Table 10.3 shows the relationship between those experiencing repeat homelessness (Repeat EHC), Once-only homelessness (Once-only EHC) together with evidence of mental health, drug- or alcohol-related issues. This is compared to their respective controls in the MDC and LDC, matched on age and sex.
The Repeat EHC has fewer people with no of drug-, alcohol-related or mental health issues
39% of the Repeat EHC have no evidence of drug-, alcohol-related or mental health issues, compared with 74% for MDC and 87% for the LDC. This is lower than the 55% of the Once-only EHC who have none (compared with 74% for the MDC and 86% for the LDC). The proportions for the controls being similar between the once-only and repeat, suggests that the difference between once-only and repeat is not due to the difference in the age distribution.
The Repeat EHC has more people with all of drugs, alcohol or mental health
Over a tenth (11.4%) of people in Repeat EHC have evidence of all three of mental health, drug- and alcohol-related conditions, compared with 1.0% for MDC and 0.2% for the LDC. This is higher than the 3.8% of the Once-only EHC who have all three (compared with 1.0% for the MDC and 0.2% for the LDC). Even though the proportions among the controls are similar between the once-only and repeat groups, the proportion for the Repeat EHC is around three times larger than the proportion for the Once-only EHC. Although the Repeat EHC and Once-only EHC have different age structures, the difference here is too large to be explained by age differences alone (see Section 2.3.2). There is a relationship between repeat homelessness and drug-, alcohol-related and mental-health issues.
Table 10.3: Proportion of people with drug-, alcohol-related or mental health issues, by cohort and repeat or once-only homelessness.
|Number of People in the Cohort||316,067||316,067||316,067||119,786||119,786||119,786|
|Any mental health||43.6%||25.8%||13.7%||59.4%||25.3%||13.2%|
|Alcohol, drugs and mental health||3.8%||1.0%||0.2%||11.4%||1.0%||0.2%|
10.5 Drug, alcohol and mental health Issues amongst different groups
This section considers how those in particular groups are affected by mental health, drugs and alcohol issues. The groups focussed on here are:
- those who were previously a member of the armed forces,
- whether someone has slept rough in the three months preceding their homelessness application
- whether someone has previously been looked after as a child by a local authority
- whether someone has become homeless immediately following discharge from prison
- whether someone has experienced domestic abuse or violence.
These groups have been chosen to suggest areas of further analysis, rather than provide definite proportions. For example, splitting the EHC into those that have been discharged from prison may result in forming groups which have different age-sex balances. The differences between these and the people in the EHC seen may therefore not just be due to whether the person has been discharged from prison, but also in differences in the age-sex structure of each group.
The Scottish Government HL1 datasets does not identify which individual in the homelessness application has these attributes. Thus if one person in the application is in one of these groups, then all household members are assumed to be in group. As most homelessness applications contain only one person, this issue is unlikely to have a big impact on the resulting analysis.
For this section of the study, the homeless data was aggregated across all assessments to analyse the above groups. For example, for the analysis of prison leavers, the dataset was analysed to see whether a person had been recorded as homeless directly from prison across all the HL1 data provided for the study – it was not just based on the most recent homeless case.
Table 10.4: Additional analysis by groups of interest
|Groups of interests*||People with any evidence of:||Observed count||All people in this group||% none||% all three|
|Alcohol||Drugs||Mental health||People with none||People with all three|
|Previously a member of the armed forces||1,285||1,396||5,018||5,360||655||10,495||51%||6%|
|Someone who has experienced domestic abuse or violence||7,553||11,189||36,890||35,929||4,892||73,572||49%||7%|
|Someone who has previously been looked after as a child by a local authority (looked after)||2,301||3,797||7,320||4,982||1,687||12,533||40%||13%|
|Evidence of rough sleeping||9,208||13,392||24,694||11,620||6,018||36,824||32%||16%|
|Evidence of rough sleeping and having been looked after.||659||1,107||1,598||480||536||2,116||23%||25%|
|Someone has become homeless immediately following discharge from prison||5,097||9,266||11,892||2,521||3,976||14,697||17%||27%|
*Note: People may be in more than one group of interest
People who have slept rough, been formerly looked after by a local authority or discharged from prison all appear to have higher proportions of people with all three issues – drug-, alcohol-related and mental health issues.
This is not surprising. Drug use amongst prisoners and formerly looked after people is known. For example, three quarters of prisoners (75%) who were previously in care as children reported using drugs in the 12 months before coming into prison  . For the prison population, the percentage testing positive for illegal drugs when entering prison has been relatively stable since 2010/11, ranging between 70% and 77%. In 2016/17, on liberation, 30% tested positive for illegal drugs  . Evidence of substance misuse amongst people sleeping rough is also known  .
Putting aside the differences in age-sex groupings between the different groups, further work may wish to consider the degree to which these factors influence health and homelessness.
10.6 Proportion of people with drug, alcohol and mental health issues by Local Authority
For each person in the EHC cohort, it possible to conduct limited analysis by local authority. Each person in the EHC is assigned to the local authority of their most recent homelessness assessment. The proportion of people within the EHC with all three issues – drug-, alcohol-related and mental health issues - is then calculated for each local authority.
Figure 10.3 is a funnel plot which shows how this proportion varies by local authority. The dotted blue line shows the EHC average for the study (around 6%). If a local authority lies within the red funnel these are not statically different from the EHC average. Points above the upper red line are statistically higher than the EHC average and are very unlikely to have occurred by chance alone. Points below are statistically lower than the EHC average.
The reasons for the differences between local authorities may be driven by:
- An actual higher proportion of people in each local authority with drug, alcohol and mental health issues
- The age-sex distributions of homeless people between local authorities may be slightly different, driving some differences
- Different recording practices across health boards could drive some of these differences, although a number of different datasets have been used for this section.
- Greater availability of drug, alcohol and mental health services may be greater in some areas, which could cause some of this difference.
Figure 10.3: Proportion of People in the Cohort with Drug-, Alcohol-Related and Mental Health Issue by Local Authority
All local authorities above the upper red line have higher proportions of EHC people with all three issue than the EHC average. Of these, North Ayrshire and East Ayrshire are the highest. This is consistent with the Drug-Related Hospital Statistics for NHS Ayrshire and Arran. These statistics show that, after taking into account the age structure of the population, this health board has the highest rates in Scotland – typically 1.3 to 1.7 times the Scottish average (depending on the measure used)  .
Inverclyde, Aberdeen City, Argyll & Bute and Stirling are also amongst the highest. The same comparison with their health board areas is more complicated  .
For example, Stirling is in NHS Forth Valley along with Clackmannanshire and Falkirk. As Clackmannanshire and Falkirk have much lower proportions than Stirling, the overall effect will be to reduce the figure for NHS Forth Valley. From figure 10.4, if the data was combined for these authorities to get a proportion for the health board, the proportion of people might be somewhere within the red funnel or close to the EHC average. In fact the Drug-Related Hospital Statistics figures for NHS Forth Valley are very close to the Scottish average on all the measures  .
Of the local authorities below the lower red line, East Dunbartonshire, Dundee City, Aberdeenshire, Angus and Moray are amongst the local authorities with the lowest proportions of EHC people with all three issues.
North and South Lanarkshire are covered by NHS Lanarkshire. This has below average scores on two out of three measures  .
The island authorities – Eilean Siar and Shetland appear above the upper red line suggesting higher than average rates for all three conditions, compared to the EHC population. However, the Drug-Related Hospital Statistics figures for these NHS boards in these areas are below the Scottish averages on all the measures.
A majority of the EHC did not have any evidence of mental-health, drug- or alcohol-related interactions during the study period (51%). This is lower than for both control cohorts ( MDC 74%, LDC 86%).
Just under a third of the EHC (30%) have had a mental health issue which excludes drug- or alcohol-related issues. This was higher than in the control groups.
There was evidence of drug and/or alcohol-related interactions for the remaining fifth of people (19%). Of these, the vast majority (94%) also had evidence of mental health issues.
The proportion of people with all three conditions – drug-, alcohol-related and mental health issues is higher among the EHC (5.9%) than in the MDC (1.0%) or the LDC (0.2%). A higher proportion of males have all three issues than females for the EHC. The difference between males and females was generally among those aged 26+. In general the proportions of the EHC that had evidence of all three issues was higher among the age range 26–60 years. Furthermore, the proportion of people with all three issues is much higher among those that have had multiple homelessness assessments (11.4% compared with 3.8% for once-only homeless). This difference cannot be explained by the younger age profile amongst the repeat homeless cohort, suggesting a relationship between repeat homelessness and drug-, alcohol-related and mental health issues.
People who have slept rough, been formerly looked after by a local authority or discharged from prison all appear to have higher proportions of people with all three issues. This is consistent with other research.
Analysis suggests there is variation across local authorities and this variation appears to be broadly consistent with Drug-Related Hospital Statistics.