Intended and unintended consequences of the implementation of minimum unit pricing of alcohol in Scotland: a natural experiment

Component design

We employed a repeated cross-sectional natural experiment design to study the impact of introducing MUP in Scotland. The natural experiment was the introduction of MUP in Scotland and we used the north of England as the comparison group. The component involved an audit of all alcohol-related attendances at ED, anonymised administrative data provided from hospitals and a face-to-face interviewer-administrated survey to ED patients. The component methods and tools were based on previous studies of alcohol-related attendances in EDs in England110 and experiences from the Scottish Emergency Department Alcohol Audit. 111

Setting

We recruited EDs from two hospitals in Scotland (Edinburgh and Glasgow) and two in the north of England (Liverpool and Sheffield). Table 1 shows the 2019 population137 across four cities that were chosen for this component. These cities had comparable population size and population composition. Table 1 also presents the proportion of data zone (Scotland)138/lower-layer super output areas (LSOAs) (England)139 in the most deprived 10% in Scotland/England. Edinburgh and Sheffield are the less deprived areas, whereas Glasgow and Liverpool are the more deprived areas. Northern England was chosen as the comparison group because it had the most similar drinking patterns to Scotland in terms of higher levels of hazardous drinking and binge drinking, economics and culture. 140–142

Data collection took place over three 3-week waves. The baseline wave was conducted in February 2018 before the implementation of MUP on 1 May 2018. There were two post-implementation follow-ups in September/October 2018 and February 2019. In each wave, data collection took place from 20.00 until 03.30 the following day from Thursday to Sunday, and from 09.00 to 16.30 on Monday to Wednesday.

We also requested anonymised information (i.e. sex, age group and diagnoses) collected routinely on all attendees over the 3-week collection periods for each wave.

Participant selection

The target population was all patients aged ≥ 16 years who attended an ED to receive acute treatment for a health condition. Trained research nurse interviewers considered all patients aged ≥ 16 years for approach and decided whether or not to approach them based on the following exclusion criteria:

• patient too unwell

• too distressed

• grossly intoxicated (alcohol)

• grossly intoxicated (drugs)

• cognitive impairment

• police in attendance

• clear language barrier and no interpreter available

• patient already participating

• routine follow-up that has been instigated by ED staff

• patient left department

• patient admitted

• staff safety issue

• end of shift

• dead on arrival

• other.

Research nurse interviewers recorded reasons for not approaching, as well as the sex and age group of the patients who were not approached.

Potential interviewees were then given written information about the component and had up to 40 minutes to decide whether or not to take part. Face-to-face structured interviews were carried out by interviewers using iPads (Apple Inc., Cupertino, CA, USA). There was a formal screening where the potential interviewees were asked eligibility questions before consent was taken. The eligibility criteria were:

• aged ≥ 16 years

• able to speak English or interpreter available

• a new ED presentation during the shift

• conscious

• physically and mentally well enough

• sober enough (alcohol)

• sober enough (drug)

• still in the department (i.e. had not left or been admitted)

• safe for staff to approach.

Eligible interviewees were asked to sign their consent on an iPad and whether or not they further consented to linkage of their hospital notes to the interview data. For interviewees who consented to data linkage, we requested demographic characteristics (i.e. date of birth, sex and postcode), attendance details, discharge status and diagnoses for attendances. More details about reasons for not being approached, interviews being terminated and failing the inclusion criteria can be found in Appendix 2, Tables 14–16.

Outcomes

The primary outcome of interest was the proportion of alcohol-related attendances among all ED patients we recorded through either observation or interview. We used a dichotomised score, where 1 indicated that the attendance was alcohol related and 0 indicated that the attendance was not alcohol related. Attendance was considered alcohol related if the attendee was not approached because of alcohol intoxication, if the potential interviewee was ineligible because of being not sober enough due to alcohol, if the interviewee had binged (≥ 6/8 units for women/men) in the last 24 hours or if the attendee self-reported that the attendance was alcohol related because of their own or another’s drinking.

We also analysed a range of secondary outcomes that were related to alcohol-related diagnosis, alcohol use and drinking patterns. Table 2 lists all the secondary outcomes and the corresponding analytic samples. Most of the secondary outcomes were binary, unless specify otherwise.

Covariates

Our primary outcome focuses on all ED attendees who were recorded by research nurse interviewers. Information about attendees who were not interviewed was limited to reasons for not approaching/interviewing, as well as whether or not they were too intoxicated to participate. Interviewers also recorded sex and age group for unapproached attendees based on triage data or their observation. This information allowed us to adjust for sex and age group in the analysis of the primary outcome.

The anonymised data from the hospitals contained information about sex and age group of all attendees. Therefore, we adjusted for sex and age group in the analysis of alcohol-related diagnosis.

The questionnaire covered a range of sociodemographic data, including sex, age, ethnicity, first four digits of postcodes, employment status, marital status and housing ownership. Area-based deprivation scores were assigned to each respondent based on their postcodes. We used the 2011 Carstairs area deprivation scores calculated for wards in England and postcode sectors in Scotland. 145,146 This gave geographies with similarly sized populations and so a measure of deprivation comparable across all four cities and the two countries. In Scotland, postcode sectors were sometimes split between two Carstairs deciles when a postcode sector covered two councils. We used a population weighting method to assign a Carstairs score to the whole postcode dependent on the population split between the councils. These variables were used as covariates when we analysed the secondary outcomes.

Sample size

As stated in our published protocol:39

. . . based on the experience of the 24 hour survey of EDs in England,111 and the assumption that at least 50% of eligible ED attendees would be recruited, we anticipated that the four sites would result in 940 recruits per week. Recruiting over three three-week data collection – giving a total sample size of 5640 – would mean that we would be highly powered (> 80%) to detect an effect size of ± 5% in the proportion of alcohol-related attendances from an estimated 30% with 95% significance. We used a base rate of 30% informed by the 24 hours audit of EDs in England,111 and assumed a 5% decrease would be of public health importance and may be expected based on current evidence. For subgroup analyses, we would have good power (> 80%) to detect an effect size of 0.23 on the FAST score among those from the most deprived quintile (estimated to be 25% of attendances), and an effect size of 0.27 among those aged 18–24 (estimated to be 15% of attendances).

Reproduced with permississson from Katikireddi et al. 39 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/

Statistical analysis

We evaluated the impact of the introduction of MUP by fitting fixed-effects multivariate regression models. For our main analysis, we fitted the following models:

in which y is the outcome variable, country is a dummy variable where 0 = England and 1 = Scotland, time is another dummy variable where 0 = before the introduction of MUP and 1 = after the introduction of MUP, hospital is a series of dummy variables (0 = Edinburgh, 1 = Glasgow, 2 = Liverpool and 3 = Sheffield), wave is a series of dummy variables (0 = wave 1, 1 = wave 2 and 3 = wave 3), MUP is the interaction term between country and time (i.e. a dichotomous indicator with the value 1 for patients who attended Scottish EDs after the implementation of MUP and 0 otherwise) and ε is the residuals. Our coefficient of interest is β₁, the DiD estimate, which is defined as the difference in outcome between Scotland and England before and after the MUP was introduced in Scotland. We used logistic regression for binary outcomes and linear regression for continuous outcome.

Model 1 is the unadjusted model with the DiD estimate and fixed effects for country and time. The country fixed effects control for all unobserved country-specific factors that are time invariant, whereas the time fixed effects account for all seasonal effects over time. In model 2, we further adjusted for hospital and wave fixed effects. As the country and time fixed effects in the unadjusted model were confounded with the newly included hospital and wave fixed effects, we omitted them from model 2. In the final model, model 3, we further included a set of covariates (i.e. sex, age group, ethnicity, employment status, marital status, housing ownership and Carstairs deprivation score). We also performed stratified analysis to examine the primary and secondary outcomes by sex, age group, ethnicity, employment status, marital status and housing ownership.

Appendix 2 provides the percentage of missing data of each demographic (see Appendix 2, Tables 18–20) and outcome variable (see Appendix 2, Tables 21–25) by country and wave. We imputed all variables in the data set (except the anonymised data set requested from hospitals) using multiple imputation. A total of 20 imputed data sets were created and subsequently analysed in R using the MICE (multivariate imputation via chained equations) package (version 3.70) (The R Foundation for Statistical Computing, Vienna, Austria). 147 The parameters from the models were estimated in each imputed data set separately and combined using Rubin’s rules.

We included non-response weights in the imputation process and analysis regression models. Using the anonymised information for all attendees from the hospitals, we were able to calculate inverse probability weights to adjust for the differences in distribution of sex and age group between attendees and interviewees.

We undertook various sensitivity analyses to investigate whether or not our results were sensitive to the model specification and the selection bias of the sample. To examine whether or not our findings were sensitive to the FAST cut-off score, we analysed the effect of MUP against FAST cut-off points of ≥ 2 (hazardous drinker), ≥ 4 (harmful drinker) and ≥ 6 (dependent drinker). These cut-off points were validated using data from the Adult Psychiatric Morbidity Survey 2007. 148 We replicated the analyses on alcohol-related attendance (primary outcome) and alcohol-related diagnosis (secondary outcome) using the sample based on interviewees and by including ethnicity, employment status, marital status, housing ownership and Carstairs deprivation score as covariates. Finally, we performed the weighted and unweighted analyses on the complete cases, which included only cases with no missing data on both outcome variables and covariates.

Training

A single-site first pilot was carried out and each individual site had a further pilot. Each study site underwent a similar training session before each wave, and data collection times sites and methods were confirmed at baseline and reiterated in the training for subsequent waves. At the EDs, there was a 3- to 4-hour face-to-face training session. The nurses who had been recruited to conduct the data collection underwent the training. It was occasionally challenging to have all staff present for the training because of absence or if staff were required to keep the service running while the training was delivered. Where staff could not attend the face-to-face training, the lead nurse was given the training presentation to cascade to any staff who were unable to attend. Training was very interactive, consisting of a 15- to 20-minute presentation, outlining the aims of the research and an overview of the research tools, followed by a hands-on opportunity to use the iPads/paper questionnaires, followed by role play completion of the questionnaire, with certain scenarios presented by the trainers.

During the run up to baseline data collection, we held informal ‘huddles’ with ED sites by telephone conference to allow for the discussion of informal queries. These were useful both in resolving queries and keeping relationships with sites healthy.

Data collection times

The component protocol stipulated that all sites should collect data at the same time. ED data collection was from 09.00 to 16.30 Monday to Wednesday and from 20.00 until 03.30 from Thursday to Sunday over three 3-week periods. We could not go outside these times, but a 30-minute period after each shift was allowed for the completion of already started cases.

Research tools

Each site was given the same mechanism for data capture. This involved purchasing iPads for each ED site, making sure that all iPads were updated with the same version of iOS, that they were connected to the King’s College London (London, UK) server and that they had the same version of the data collection application (app) and all were able to access the internet through Wi-Fi within the site or through 4G mobile data. It was also imperative that each site updated the data collection software at the start of each wave of data collection to ensure that any data collection app updates were implemented and that all sites were working from the same version of the app. Sites were asked to manage the use of their iPads, ensuring that they were charged and data synchronised at the end of each shift. Although this was complex to administer across sites, we were able to synchronise the iPads successfully and this challenge was overcome.

Maintaining similar and adequate staffing levels

Given the busy nature of EDs, and we always seemed choose busy periods, it was difficult for the lead nurse at each site to schedule nurse shifts for the duration of the data collection. It was recommended that each shift had three members of staff allocated. Two research nurses would interview participants and one team member would provide administrative support. There were slightly differing models of utilising the staff, depending on the requirements of each ED. For example, in Edinburgh, one administrator was dedicated to monitoring which patients were eligible for interview, with two nurses conducting the interviews. In other sites, three nurses were used to conduct interviews. It was appropriate that each site was able to manage this individually, as the physical space and triage systems were different in each ED and it was felt that local sites were best placed to decide how to manage this aspect of data collection. However, consistency across waves was achieved by using a similar delivery model at each wave and regular weekly catch-ups with sites during each wave.

Confidentiality

The physical space in each ED was different, but a common issue reported by nurses was maintaining confidentiality during the interviews. Where possible, the interview was conducted in a private room; however, this was not always possible. Nurses were advised to respect patient confidentiality at all times, in line with their standard practice. This meant not conducting interviews within earshot of other patients or family members. This not only avoided breaches of confidentiality, but also encouraged interviewees to answer freely and honestly without fear of judgement from others.

Missing participants

On particularly busy shifts, there was potential to ‘miss’ patients who were in the ED for a short period of time. Nurses minimised this by monitoring patient flow on the ED computer system and approaching them at the earliest appropriate time.

Language barriers

It was difficult to ensure that there were always interpretation services available. Staff were trained to make use of interpretation services (i.e. a face-to-face or telephone-based interpreter) where possible; however, nurses informed the research team that this was often not possible as it could take time to connect with an interpreter, which would lead to missing other participants. Nurses were instructed not to make use of participants’ family members to translate, so as not to breach the participants’ confidentiality. There were very few cases recorded as not being able to participate because of language issues.

Avoiding disruption to regular workings of the emergency departments

Throughout the data collection process, it was essential to avoid disruption to the work of the ED for participants and other patients who were in attendance at the ED. Nurses reported that they were occasionally asked to contribute to non-research-related tasks and the principal investigator for each site was encouraged to instruct ED colleagues that the research nurses were not to be asked to complete non-research-related work unless it was completely necessary.

Obtaining sufficient recruitment numbers

Eligibility criteria meant that a number of staff could not be interviewed, but we collected observational data to measure our primary outcome. It was often a chaotic data collection environment; however, as a condition of ethics approval, we did not want to disrupt normal care for patients and so interviews were terminated while in process if a patient was to be urgently admitted.

Contextual issues

The ‘Beast from the East’ snowstorm closed data collection at most sites on 28 February 2018 and that day was excluded for all sites. There were also some staffing issues with, for example, staff off sick. Generally, the timeline for preparation for collecting data was short because of late announcement of the legality and then the relatively swift implementation date.

Primary outcome

Figure 2a shows the changes in the proportion of recorded attendees with alcohol-related attendance in Scotland and England before and after the introduction of the MUP. On average, Scotland had a higher proportion of attendances that were alcohol related than England. Scotland had a stable trend, whereas there was a decreasing trend in England. There was no evidence of statistically significant differences in the primary outcome after the introduction of MUP in Scotland. The odds ratio (OR) of having an alcohol-related attendance before and after implementation of the MUP was 1.14 [95% confidence interval (CI) 0.90 to 1.44; p = 0.272] in Scotland compared with England (see Figure 4 and Appendix 2, Table 26, for the summary of DiD estimates from full regression models).

FIGURE 2.

Outcome trend across three waves compared between Scotland and England. (a) Alcohol-related attendance; (b) alcohol-related diagnosis; (c) current alcohol drinker; (d) binge drinking in past week; (e) binge drinking in past 24 hours; (f) FAST score; (g) alcohol misuse (FAST score ≥ 3); (h) increased alcohol use in past year; (i) place of last drink (private location); and (j) place of last drink (on-licensed premise).

Secondary outcome

In contrast to alcohol-related attendance, which was lower in England, England had a higher prevalence of alcohol-related diagnosis than Scotland (see Figure 2b). Note that our comparison was with hospitals in the north of England. The proportion of attendees with at least one alcohol-related condition rose slightly in Scotland, but fell in England. Figure 3 shows the changes in proportion of different alcohol-related diagnosis across waves. There were more attendees diagnosed with partially chronic alcohol-related diagnoses among all alcohol-related diagnoses, followed by wholly acute, partially acute and wholly chronic alcohol-related diagnoses. As the prevalence of partially acute and wholly chronic conditions was very low, we combined all alcohol-related diagnoses as one outcome in the analysis.

FIGURE 3.

Trend for alcohol-related diagnosis by conditions across three waves. (a) Wholly chronic alcohol-related diagnosis; (b) partially chronic alcohol-related diagnosis; (c) wholly acute alcohol-related diagnosis; and (d) partially acute alcohol-related diagnosis.

Across waves, there was a slightly increased trend in being a current alcohol drinker in both countries (see Figure 2c). Binge drinking in the past week among all respondents increased slightly in Scotland, but decreased in England (see Figure 2d). However, both countries showed a slight increase in binge drinking in the past 24 hours across waves (see Figure 2e). The mean FAST score among drinkers increased in both Scotland and England (see Figure 2f). The proportion of alcohol misuse (i.e. a FAST score ≥ 3) increased in England, whereas Scotland had a relatively stable trend (see Figure 2g). Meanwhile, the proportion of drinkers who reported an increase in alcohol use in the past 12 months also had a stable trend in both countries (see Figure 2h). Figures 2i and 2j show that the proportion of drinkers who had their last drink in a private location was higher than those in on-licensed premise. However, the trends were stable for both private location and on-licensed premise across waves.

Figure 4 shows the DiD estimates from the final regression models for our secondary outcomes (see Appendix 2, Table 26, for the summary of DiD estimates from full regression models). There was no evidence of significant differences in most outcomes after the introduction of MUP in Scotland. The DiD estimates show that among all attendees the odds for an attendee having at least one alcohol-related diagnosis increased by 25% relative to change observed in England after MUP (OR 1.25, 95% CI 1.00 to 1.57; p = 0.046). Nevertheless, there was no effect on other secondary outcomes, suggesting that the introduction of MUP in Scotland did not substantially alter these outcomes in the population studied.

FIGURE 4.

Difference-in-difference estimates of the overall effects of MUP (a) based on all recorded attendees; and (b) based on drinkers only.

Primary outcome

Figure 5 shows the stratified results for the primary outcome. There was no evidence to show that MUP had any differential effect across sex and age group. The summary of DiD estimates from the full regression models were presented in Appendix 2, Tables 27–39).

FIGURE 5.

Stratified analysis for primary outcome: alcohol-related attendance. Note that the p-values in the forest plot were uncorrected.

Secondary outcome

Glasgow

• The 20-minute integrated clinic, test only, nurse return clinic, follow-up clinic, vaccine clinic, gynaecology clinic and also menopause and contraceptive procedures.

• Times: 09.00–17.00, Monday to Friday (as per protocol).

• Researchers: clinic reception staff (waves 1 and 2) and research nurses from SPHSU bank (wave 3).

Edinburgh

• All clinics.

• Times: 09.00–17.00, Monday to Friday (as per protocol).

• Researchers: clinic reception staff (all waves).

Dundee

• All clinics (walk-in and booked).

• Times: 09.00–17.00, Monday to Friday (as per protocol).

• Researchers: clinic reception staff (all waves).

Sheffield

• All clinics (including youth clinic to 18.00).

• Times: working hours between 09.00 and 17.00, Monday to Friday (as per protocol). Wednesday morning only (not Wednesday afternoon). Urgent care, nurse return clinic, vaccine clinic, pregnancy clinic, and menopause and contraceptive procedures. Not for self-tests or for people making an appointment.

• Researchers: research nurses.

Manchester

• All clinics: human immunodeficiency virus (HIV), family planning and contraception clinics, and the general genitourinary medicine walk-in.

• Times: the time slots were from 08.00 to 16.00 with walk-in sessions from 08.00 to 15.00.

• Researchers: research nurses.

Leeds

• All clinics: booked in appointments and walk-in attendances.

• Times: 09.00–17.00, although this was changed to 11.00–17.00 for team brief days.

• Researchers: research nurses.

These sites, times and staff were replicated across waves, apart from a change in research staff in wave 3 in Glasgow, which was essential to improve the response rate.

Contraception patients were included at all sites. If the service design was a hub-and-spoke structure, then sites were advised that there should be enough patients at the hub and so not to include spokes. Sites were told that it was acceptable to give assistance to people who need help to complete the questionnaire. We informed sites that we would be asking for a report from the clinic information system, which would be held on computer, of the numbers of patients by age and sex who attended during the data collection time periods. Only if the clinic information system was unable to provide this information would they need to keep a manual tally of refusals. Clinics were given flexibility to vary time slots by site to allocate around 7.5 hours within the normal working hours of the site for recruitment, but it was stipulated that these should be kept the same for waves 2 and 3. People who came in to the clinic but were sent away (e.g. with a chlamydia testing kit) and were not registered as attendees were not eligible to participate.

Study size

We carried out an a priori power calculation to establish our intended minimum sample size. We calculated that a total sample of 5000 participants for each wave would detect a change of ± 4% in our primary outcome (therefore, from 30% to 34%, where 30% of people are using drugs), with 95% significance at 80% power. Our main results compare the pre-MUP baseline with all post-MUP waves, but we also compare each follow-up wave with the baseline separately to check trend.

Demographic variables: predictors, potential confounders and effect modifiers

We measured several predictors and potential effect modifiers, including age group, sex, highest educational level and employment status. The DiD design accounts for time-invariant confounders.

Comparisons

We used DiD OR for outcomes based on proportions and an estimate of the mean for the FAST score, and tested for potential effect modifiers. We present results with 95% CIs and associated p-values. We combined the two post-MUP waves in the time variable in the model by coding the baseline as time = 0 and waves 2 and 3 as time = 1.

Description of statistical methods, including those used to control for confounding

We used a fixed-effects regression model, with individuals nested within SHCs, before and after adjustment for relevant covariates. All outcomes except the FAST score are binary. We use logistic regression for binary outcomes and linear regression for the FAST score. Model 1 is unadjusted. Model 2 is adjusted for clinic site and wave. Model 3 is adjusted for clinic site, wave, age, sex, education, employment status, the interaction between age and wave, and the interaction between sex.

where y is the outcome, β₁ is the interaction term between country (0 = England; 1 = Scotland) and time (0 = before MUP; 1 = after MUP, waves 2 and 3 combined), and ε is the error term. β₁ is the DiD estimate and the reference category is ‘England before MUP’. It is defined as the differences in outcome between Scotland and England before and after the implementation of MUP. The inclusion of wave (with values of 1, 2 or 3) in the model controls for seasonal differences. The clinic variable is a series of five dummy variables and wave is a series of two dummy variables.

Description of statistical methods used to examine subgroups and interactions

We tested for statistical interactions of intervention and covariates on a multiplicative scale (including age, sex, highest educational category and employment status) to investigate the possibility of differential intervention effects, and subsequently stratified the analyses. In our regression model, we compare pre MUP and post MUP to allow for seasonal effects or if the effect of MUP is not immediate.

Missingness for demographic variables was between 0% and 13% when assessed for waves within countries. Missing data for both demographic and outcome variables were addressed through 20 rounds of multiple imputation using Rubin’s method and chained equations using the MICE package147 in R. We applied weights in our regression analysis according to age group within each clinic to ensure that our sample was representative of age for all clinic attendees.

Our sensitivity analysis compared outcomes using complete-case analysis and using multiple imputation. We found only minimal differences (see Appendix 3, Tables 64 and 65). We compared weighted with unweighted samples, finding minimal difference. Figures showing these analyses are included in Appendix 3, Figures 29–38. We also carried out an analysis of outcomes excluding the clinic for which we introduced extra support at the final wave (see Appendix 3, Table 66). The findings were not affected. Our final results are for model 3 after imputation and weighting.

Illicit drugs results

Alcohol outcomes

Anticipated outcomes

Professional stakeholders representing each study region and a diverse range of categories expressed expectations that MUP would reduce purchasing and consumption of alcohol. One youth worker explained how rising prices, in the context of finite household budgets, would reduce purchasing:

Yes, because they’ll know how much money they’ve got and when you add it onto your shopping list that takes you over. You’ve still got gas and electricity. So, I do think, I do think in that respect it’s quite a good thing.

Baseline, affluent West, youth worker

However, the same participant went on to suggest that these positive effects may unfold gradually because of price increases being ‘not massive’. Despite participants identifying limited budgets as a factor in the mechanism of MUP, there was limited overt discussion of MUP affecting purchasing and consumption among low-socioeconomic-status groups.

Stakeholders predicted reductions in alcohol consumption for specific social groups, most notably young people (implicitly teenagers), constructed as constrained by limited budgets and able to make rational decisions based on budgetary pressures. A youth worker explained that:

If it’s not accessible and cheap enough there will be a realisation that they’ll manage it. [. . .] Because they do have limited resources.

Baseline, deprived West, youth worker

A general practitioner (GP) described how the measure could have an impact on young people:

I mean the other group I think it may have an impact on, which isn’t specifically in a deprived population, is younger people who’ve got a relatively limited disposable income.

Baseline, deprived East, GP

A stakeholder working in addiction services suggested that young people will be driven to opt out of some episodes of drinking:

A lot of these young people will maybe be able to afford it sometimes, you know, depending on what they’re drinking, but it probably will have, they will probably need to look at well if we’re going to do that something else will need to suffer this week or this month or whatever.

Baseline, deprived West, stakeholder

In addition to young drinkers, some stakeholders identified different age groups as likely to be affected by MUP. One youth worker (baseline, affluent West) suggested that MUP would mean that older people ‘won’t drink as much alone in the house’, whereas another youth worker suggested that middle-aged people might question and change their own drinking behaviours:

There will be an impact and it’s how they manage that impact and sometimes for the middle aged person that’s working it’s the catalyst that makes them think a wee bit.

Baseline, deprived West, youth worker

This participant went on to suggest that middle-aged people may, instead of drinking, ‘go and talk to my pal over a cup of tea, know what I mean?’, and suggested that such people’s ability to make that rational choice was because of their lifestyles not being ‘chaotic’. This participant identifies a social group who may engage in harmful drinking, but not to a chaotic degree, and whose behaviour may, therefore, be moderated by price increases. Similarly, a licensing authority stakeholder described:

Sensible-type families who work to a tight budget will lose a bottle of wine at the weekend.

Baseline, affluent West, licensing authority

Some participants expressed doubt about the impact of MUP on consumption, particularly for more affluent groups. An addiction services stakeholder in an affluent area explained that:

The more affluent someone is, it’s difficult to get them to make changes because it’s not affecting them financially.

Baseline, affluent West, addiction services stakeholder

Similarly, a GP in the same area explained that:

I don’t think it is going to make much difference locally because I don’t think people are buying the £0.20 cans of lager to drink here.

Baseline, affluent West, GP

The same GP went on to suggest that health and family pressure, not price, motivate behaviour change in their patients:

I’ve not had anyone where the driver for help was, I can’t afford this anymore. It’s that they are becoming aware of the damage it has done. Usually not the damage it’s doing to themselves, physical health wise, it’s the damage in the family.

Baseline, affluent West, GP

Some stakeholders suggested that people in more challenging social circumstances were unlikely to limit their consumption. A police stakeholder explained that:

There’s quite a majority of them who are not that poor so it’s not like they’re all unemployed, a lot of them are older and their partner’s died and they’ve had mental health issues but they’ve got some sort of income, a pension or something. So, I don’t see them stopping because it’s going to cost them an extra pound to buy a bottle of vodka or whatever.

Baseline, deprived East, police stakeholder

One GP suggested that:

It’s going to make people drink better quality alcohol, you know, they won’t drink the really cheap stuff . . .

Baseline, deprived East, GP

They also explained that:

There isn’t really any evidence that it’s better quality, it’s just marketing.

Baseline, deprived East, GP

A range of participants covering each stakeholder category and study area predicted that people simply would not let price increases affect their alcohol use. Typifying this, a police stakeholder stated:

Folks are still going to drink no matter what your price is, by hook or crook they’ll drink.

Baseline, deprived West, police stakeholder

Although this assumption was predominantly applied to dependent drinkers, a youth worker suggested the same of young people:

If they’re drinking a £2 bottle of cider or they’re drinking a £15 bottle of vodka they’ll drink, know what I mean, so they’ll adapt their budget to whatever it goes.

Baseline, deprived West, youth worker

Some expressed related concerns that prices of certain categories of alcoholic beverage would not change substantially. A GP identified wine and associated it with women’s excessive drinking:

So certainly a lot of women are drinking wine to excess. Not to excess on a particular day to the extent of getting drunk but drinking wine regularly that tots up. So those consumers may not be affected because the individual units of bottles may not go up that much.

Baseline, deprived East, GP

A police stakeholder discussing underage drinkers stated that:

If you’re telling me that [the price of] Mad Dog [Mogen David Wine Co., Westfield, NY, USA], Dragon Soop [Corinthian Brands Ltd, Knaresborough, UK] and Buckfast [J Chandler & Co (Buckfast) Ltd, Andover, UK] won’t be affected no. [MUP will have no effect] at all.

Baseline, deprived West, police stakeholder

This police stakeholder went on to suggest that ‘hardened drinkers’ may also move to those categories of drinks.

Regarding retailer compliance, a group of licensing authority stakeholders (baseline, deprived East) observed that, although price marking can be monitored, informal discounts given by shopkeepers to regular customers cannot and, therefore, some shops may not comply wholly with the minimum price.

Observed outcomes

Licensing authority stakeholders described predominantly positive observations of retailer compliance, both before and after implementation. Prior to implementation, one stakeholder reported that a supermarket had already adjusted their pricing and suggested that compliance was likely to be widespread:

I’d taken a stroll down to Asda [Asda Stores Ltd, Leeds, UK], our supermarket here, and the likes of the one in [nearby affluent area], they’re probably already roughly compliant.

Baseline, affluent West, licensing authority

Subsequent to implementation, one licensing authority stakeholder reported that compliance was:

Up at 90-odd per cent, at least, of full compliance.

Follow-up, deprived West, licensing authority

Whereas another reported that:

I’ve had no issues at all with any of the shopkeepers I’ve dealt with.

Follow-up, deprived East, licensing authority

Although licensing stakeholders’ observations were broadly positive, there were some anecdotal exceptions. One such stakeholder described observing a situation in which the manager of a branch of a chain based in England was put under pressure by having to comply with MUP when selling stock from suppliers in England:

The likes of her, for the first couple of weeks, you are talking 60% of her stock . . . she was having to calculate it and relabel it all. So, for the first month or so, maybe up to 2 months it was quite horrendous. But then, I don’t know, it seemed to calm down.

Follow-up, affluent West, licensing authority

This observation characterises MUP as a substantial operational inconvenience for some retailers, but a temporary one to which they can adjust. Another, licensing authority stakeholder perceived MUP as having increased small retailers’ competitiveness with supermarkets, but identified the potential – but not yet observed – unintended consequence of causing the re-emergence of an overprovision problem that had previously ceased to be relevant. The participants explained that:

People are using their local shops more often, or they certainly appear to be because there’s no price difference, it’s made it more relevant again because part of the thing before was most people wouldn’t go to their local shop and pay £3 or £4 extra for a bottle of vodka, whereas now because they are right across the board, if anything it’s made overprovision more relevant again because you know the prices are uniform.

Follow-up, deprived East, licensing authority

A limited number of anecdotal observations from stakeholders suggested that purchasing and consumption of alcohol had reduced. An addiction services stakeholder described a perceived reduction in consumption of white ciders, reporting that ‘I’ve not seen anybody down as drinking the cheap stuff’ and going on to say ‘It does get people thinking, “oh right”, maybe on a tighter budget’.

The same stakeholder described one woman who replaced vermouth with mulled wine. A youth worker identified a reduction in young people’s consumption of spirits, observing that:

The cheaper options are [now] more expensive options and I think the spirit thing – it’s been great cos it’s not been as easy to access. Even though you club together, you still cannae really get good value, especially if there are six or seven of you with one bottle.

Follow-up, deprived West, youth worker

A youth worker described young people’s self-reported adaptations to changing affordability:

It’s really interesting conversations about how they self-manage, to still have alcohol as part of their routine, which is really interesting, because of the good weather, they were drinking every night but they were still able to self-manage. Likely they had the same alcohol intake over the 5 days, but they wouldn’t finish drink because they were sharing it every day. I was saying that is a wee bit sensible. They were kind of taking their units in a more sensible way, rather than binge drinking.

Follow-up, deprived West, youth worker

This account also highlights how contextual factors, such as an increased drive to consume alcohol in warm weather, may moderate the impact of MUP.

A substantial set of participants described having observed no evidence that MUP affected purchasing or consumption of alcohol, and many participants characterised drinkers as resilient and resistant to change. Describing shopkeepers’ experiences, a licensing authority stakeholder explained that:

Nobody’s really said they’ve noticed a change in trends. That they could come out and say, oh [name], I’m now selling much more of – everybody that was drinking high strength cider has now moved on to vodka kind of thing. People are just saying they are selling about the same.

Follow-up, affluent West, licensing authority

The same licensing authority stakeholder as above explained that:

If you are a Frosty Jack [Aston Manor, Birmingham, UK] drinker, you are still a Frosty Jack drinker. You are just paying more for less.

Follow-up, affluent West, licensing authority

Some youth workers perceived MUP as having had little effect on young people’s consumption. One youth worker observed that:

[MUP] doesn’t seem to impact on [young people’s] ability to purchase it and consume it and they use it as part of their everyday life.

Follow-up, deprived West, youth worker

Another youth worker described MUP as not deterring young people because:

Some of them were drinking underage as well, so their ability to get alcohol cheap and illegally – they could do that anyway. It hasn’t stopped any of the older ones buying alcohol or using it.

Follow-up, affluent West, licensing authority

The same youth worker suggested that wine drinkers may not have been deterred either, because:

They can still go to Aldi [ALDI Einkauf GmbH & Co. oHG, Essen, Germany] and get a bottle for £4.98.

Follow-up, affluent West, licensing authority

Anticipated outcomes

Participants from each study region, and a diverse range of stakeholder categories, expected MUP to produce improvements on population health over the long term. An addiction services worker typified this, stating:

Longer term, if the population as a whole was drinking slightly less, it’s bound to be more beneficial, physically and mentally.

Baseline, affluent West, addiction services worker

Some stakeholders indicated that improvements in health would be expressed in fewer presentations at EDs with alcohol-related illness, whereas others identified a risk of a short-term increase in presentations due to withdrawal. One GP (baseline, affluent West) characterised these as ‘short-term potential health risks’ that ‘won’t be able to go on indefinitely once people get used to the new pricing’. Similarly, an addiction services stakeholder explained that:

It’s only going to do good, but until we get to the good there’s going to be harm and other things on the way.

Baseline, deprived West, addiction services stakeholder

Some participants anticipated that health improvements would not be distributed equitably across the population. A GP (baseline, affluent West) identified that benefits will be experienced by ‘a very particular group of people with a certain income bracket, a certain amount of intake’, going on to specifically identify heavy drinkers in deprived areas. The same participant also predicted that some drinkers would manage to reduce consumption without experiencing withdrawals, such that they receive health benefits that will not be detected by health services. Some participants envisaged potential reductions in alcohol-related injury. A youth worker (baseline, affluent West) identified ‘people in their 30s and 40s who just routinely drink more than they should’ as a group that may be less likely to report at accident and emergency with injuries. Although participants from each stakeholder category expressed positive expectations for purchasing, consumption or health, only one of the six police stakeholders expressed such expectations, and did so reservedly:

Well obviously the policy is there to try and improve the health and lessen the intake of alcohol. So, I mean I’m hopeful, hopefully it will.

Baseline, deprived West, inspector police licensing

Some participants identified the magnitude of the reduction in intake they perceived necessary for some drinkers to conform with safe drinking guidelines. An addictions services worker explained that:

From 200 to 150 [weekly units] is 25%. So that is a big reduction. Yes, if you did have someone present to services and was able to – if you could get that sort of reduction, you would see some health benefits.

Baseline, affluent West, addiction services worker

One GP explained that:

Even if they reduce [consumption] by 20%, they are still going to be well over the recommended limit . . . I can’t see it making enough difference. To get someone from 50 units a day to 2 units a day, I don’t know how the minimum pricing is going to affect that.

Baseline, affluent West, GP

Although some participants characterised increased presentation at health and support services as a short-term cost for long-term benefits, others expressed more concern. An addiction services stakeholder anticipated increased presentation by deprived, alcohol-dependent groups:

You need to be prepared for the vulnerable group who are on very low incomes and who require to buy alcohol on a daily basis to maintain . . . there’s a huge population out there that don’t engage in services and that for me is maybe that hidden population that we have to be prepared for.

Baseline, deprived West, addiction services stakeholder

Similarly, a GP explained that:

All of a sudden they can only afford half the amount of alcohol they were drinking, so they there might be an abrupt presentation.

Baseline, affluent West, GP

Some participants characterised dependent drinkers as unable to sufficiently moderate their daily intake to ensure uninterrupted supply. An addiction services stakeholder explained that:

If you are used to say 150 – if you are drinking about 20 units a day, you’ll still drink 20 units a day. But if you can only afford for 5 days instead of the 7.

Baseline, affluent West, addiction services stakeholder

The same stakeholder predicted the re-emergence of a problematic pattern of behaviour:

When I first started working with problem drinkers, people would run out of money and for a good number of years now, that doesn’t seem to happen the same, where people would go into unplanned withdrawals, quite serious withdrawals, because of a shortage of money. I can see where that situation would come back into play.

Baseline, affluent West, addiction services stakeholder

In addition to withdrawals, a minority of stakeholders anticipated other health harms. An addiction services stakeholder identified risks to mental well-being among professionals:

Quite often I would relate it to your professional group and [alcohol] is their de-stressor. So, stress and anxiety levels could possibly increase if someone doesn’t find alternatives to using alcohol.

Baseline, affluent West, addiction services stakeholder

Observed outcomes

A small number of participants described observing anecdotal indications of improvements in health that may have been precipitated by the intervention. An addiction services worker reported that:

I’m no seeing just as much assessments coming in where it’s somebody who’s a bottle of wine a night. Speculating that: I would imagine it is that they are buying less but still sufficient to avoid unintentional withdrawals.

Follow-up, affluent West, addiction services stakeholder

The same stakeholder reported that ‘there has only been a handful I have heard from the team, so and so was admitted to hospital with withdrawals’, suggesting that withdrawals had not increased substantially subsequent to the intervention.

Although stakeholders predominantly reported no negative impact on health, one primary care stakeholder (follow-up, deprived East) expressed concern that their patients who ‘have had problematic issues with spirits and feel that part of their recovery was moving away from that’ have begun to consume spirits and ‘feel forced into doing so’ by the government.

Many stakeholders, covering each stakeholder category and study region, reported having observed neither a positive nor a negative impact on health. Some participants in primary care roles described noticing no changes in health or help-seeking. Typifying this, an addiction services stakeholder stated:

We haven’t sort of seen a massive increase, a crisis, any more than there has certainly been in my time in the team.

Follow-up, affluent West, addiction services stakeholder

Similarly, a GP stated:

I can’t say I’ve noticed any great difference (aside from) a couple of people [who] came in saying that this would make difference to them.

Follow-up, deprived West, GP

However, the same GP identified that changes may be evident only in admissions data:

It hasn’t made a major impact, but maybe on an individual practice level it is hard to see. You’ll be able to look at A&E [accident and emergency] admissions. When we are doing our day to day work, we are too busy seeing patients to almost even collate what we’re seeing.

Follow-up, deprived West, GP

Anticipated outcomes

Stakeholders from the primary care, addiction services and licensing authority categories identified potential social benefits of MUP related to crime, antisocial behaviour and domestic issues. Like discussion of impacts on consumption and health, police stakeholders did speculate about benefits in crime and social problems. An addiction services worker identified a broad range of benefits:

Well I suppose impact on children, that whole poverty thing and families, children affected by alcohol use. Criminality, crime, violence, intoxication, young people fighting, stabbing . . .

Baseline, deprived West, addiction services worker

Taking a broader view, a youth worker (baseline, affluent West) suggested that ‘Part of the policy is about trying to get the industry to be a bit more responsible’ and going on to suggest that the alcohol industry may play a role in improving labelling and public understandings of the content of alcoholic drinks.

Regarding families and domestic violence, a licencing authority stakeholder expressed cautious optimism:

If it works it will have a massive positive effect on that, but time will tell I think.

Baseline, affluent West, licencing authority stakeholder

Outside the domestic context, a GP talked about how people going out to drink less on weekends:

Maybe reduces the alcohol-fuelled crime if there is fewer people out on those nights of the week.

Baseline, affluent West, GP

They suggested that:

[MUP] might reduce how much alcohol someone can afford to buy in one day and therefore reduce drink related crimes.

Baseline, affluent West, GP

A GP foresaw potential reductions in unintended sexual encounters in young people:

They’re getting into fights or, you know, unintended sexual encounters that leave them with infections or pregnancy or what have you. So, it would be good if it had a positive impact on those groups because they tend to go out with a finite amount of money on a night out and once it’s gone it’s gone.

Baseline, deprived East, GP

The same participant suggested that young people may stay in education more effectively because of reduced alcohol intake:

We come across things like [young people have] had a college place and they’ve lost a college place because they’ve not been getting there on time after the weekends and things. So, they lose their place and they’re back on benefits, all that sort of thing. So that group it might have an impact on.

Baseline, deprived East, GP

Similarly, an addiction services worker (baseline, affluent West) predicted that reduced intake would ‘improve individuals . . . performance at work’.

Many stakeholders from the primary care, licensing authority and policy categories, and covering each of the three study areas, predicted increases in crime. Notably, stakeholders predicted that dependent drinkers might commit theft to fund their alcohol consumption. A police stakeholder explained that ‘If they’ve only got £5 and they can’t get a bottle of Frosty Jack any more they’re going to have to either go and threaten somebody or steal something to get enough money because they need that Frosty Jack’, going on to suggest that ‘they’re going to have to use violence whilst they’re alcohol fuelled to steal money or property from whoever’s house they happen to be drinking in that day’. One licensing authority stakeholder suggested that this behaviour was already taking place, but that MUP would exacerbate it:

I think they probably do run out of money on day 5. But [after MUP] they’ll run out of money on day 3. That’s not for everyone, that’s for some of them. They’ll sell or buy or steal or whatever to get alcohol.

Baseline, affluent West, licensing authority stakeholder

Some stakeholders suggested that shopkeepers would bear the brunt of customers’ anger at increased prices. Licensing authorities from both the affluent West and deprived East regions expressed this concern, with the former explaining that:

. . . a lot of my shops have contacted me and what they’re saying is they’re worried about conflict at the till. And they think there’s not been enough publicity, you know, a dependent drinker or somebody with a drink problem is going in and, you know, he’s wanting to pay £4 for his four cans of cider that he’s going to get told no that’s £9. That’s going to lead to conflict, and I don’t think these posters are going to diffuse that situation.

Baseline, affluent West, licensing authority stakeholder

A police stakeholder (baseline, deprived West) echoed these concerns, anticipating that:

Racial crimes will go through the roof because . . . the shopkeepers will get absolutely dogs abuse because it will be [seen as being] their fault.

Baseline, deprived West, police stakeholder

Displacement, whereby individuals forego food or other essentials to afford alcohol, was identified as a risk by stakeholders from addiction services, police, youth workers and, particularly, primary care, and covering each study region. One police stakeholder explained that:

All the other things in [dependent drinkers’] life are ancillary to their alcohol so they won’t eat, they won’t go and visit their family, there’s lots of things they’ll put out of the way because all they can do is this addiction they’ve got, they’ve got to fuel it.

Baseline, deprived West, police stakeholder

One GP identified that dependent drinkers’ diets are already often poor:

Some of these people barely eat. They get a lot of their calories from alcohol and they are already in a terrible situation.

Baseline, deprived East, GP

Another GP (baseline, deprived West) asked:

If you put the price up will their diet suffer even more?

Baseline, deprived West, GP

An addiction services stakeholder (baseline, deprived East) characterised the risk of displacement MUP within a context in which ‘the price of food in the supermarket has already gone up’. In addition to displacement, some identified the risk of accruing debt. An addiction services stakeholder (baseline, affluent West) suggested that ‘we’ll maybe see more people getting into rent arrears’.

An addiction services stakeholder typified concerns that displacement would also affect families:

If you think of a family on a low income and there’s a couple of kids to feed as well but somebody’s got to maintain a chronic alcohol problem and don’t at that point want to engage in services it’s going to affect the whole family.

Baseline, deprived West, addiction services stakeholder

In addition to families being neglected through displacement, a pair of addiction services stakeholders discussed the threat of domestic violence. One suggested that MUP ‘will have a massive impact [on domestic violence]’, explaining that:

We’re working with really vulnerable folk who have experienced high levels of trauma in their lives and their coping strategies maybe aren’t as healthy as, you know, others and they will react in a variety of ways that will have an impact on who else lives in that house. There’s a higher level of violence and aggression just through alcohol never mind if somebody’s under pressure.

Baseline, deprived West, addiction services stakeholder

The pair went on to identify situations where the affordability of alcohol may benefit the families of aggressive, dependent drinkers, discussing how ‘Loads of families feel safer when somebody is lying on the couch . . . Drunk, rather than when they start to resurface so in terms of aggression levels’. These accounts highlight how the risks of financial pressure and domestic violence are inter-related.

Stakeholders’ focus on how MUP would affect low-income, dependent drinkers implicitly positioned the measure as one that might exacerbate inequalities. Some stakeholders highlighted that issue more explicitly, using terms such as ‘poor’ (baseline, affluent West, GP), ‘on the breadline’ (baseline, deprived East, addiction services) and ‘at the lower end of the economic scale’ (baseline, deprived East, addiction services). Some stakeholders overtly characterised MUP as unfairly punitive of deprived groups. A police stakeholder (baseline, deprived West) described MUP as ‘a tax on the poor’, suggesting ‘That thing that makes their life a wee bit better in their eyes has now been put out of reach for them’. A GP highlighted the risk of exacerbating financial problems, explaining that:

These people are often very poor anyway and you sometimes even, without the pricing wonder where they get the money from. So, I don’t know by putting the price up will that mean people are just in more debt . . . ?

Baseline, deprived West, GP

Observed outcomes

Stakeholders predominantly did not report observing outcomes related to crime and social issues. Typifying this, an addiction services stakeholder whose work covered several deprived areas suggested that:

I don’t think there has been any increase [in financial hardship], because generally it would be myself that would sign off on a foodbank voucher.

Follow-up, affluent West, addiction services stakeholder

In contrast, a primary care stakeholder (follow-up, deprived East) perceived that, since the implementation of MUP, ‘I think my patients are accessing food banks more – these patients. I have heard – it is obviously sensitive – but they are resorting back to crime’. A youth worker had observed improvements in some people’s lives since the implementation of MUP, but pointed out the difficulty of identifying causal links:

What I will say, a couple of the people I spoke about last time, that were experiencing problematic use of alcohol. Their use has gone down and their circumstances have improved dramatically . . . I don’t think it’s been pricing that’s stopped it. But it’s hard to see.

Follow-up, affluent West, youth worker

Anticipated outcomes

Some stakeholders discussed how increased alcohol prices might drive demand for alternative substances and sources of alcohol. In one group interview, three police stakeholders discussed alternative sources of alcohol at length, cautioning that:

People who were drinking branded vodka will go to cheap vodka. People who were on cheap vodka will go to counterfeit vodka and there’s no quality control over that, they could be drinking paint thinner for all we know. So there will be unintended consequences, people will die because they’re taking uncontrolled products.

Baseline, deprived West, police stakeholders

Although these police stakeholders identified ‘black market’ alcohol as formerly associated with older drinkers, they suggested that ‘if [young people] see older people drinking this illicit alcohol at much greatly reduced prices on the black market they’ll have to look at that’.

Stakeholders identified different categories of drug abuse that may increase because of MUP. A licensing authority stakeholder posited that:

It might be cheaper to buy cannabis for example. In fact I know kids do that just now because it’s cheaper and easier to get than it is to buy alcohol.

Baseline, affluent West, licensing authority stakeholder

A youth worker (baseline, affluent West) concurred that MUP ‘will really impact on . . . increased cannabis use’, going on to explain that ‘if they don’t have their usual alcohol on a Friday down the park then, you know, there’s probably more risk of them taking other things’. One GP (baseline, affluent West) suggested that the use of ‘so-called legal highs’ may increase as a result of MUP. A GP (baseline, deprived East) suggested that dependent drinkers would not substitute alcohol for drugs, but ‘younger people’ engaged in ‘the party culture’ may do.

In addition to recreational drugs, some stakeholders discussed prescription drug abuse. One addiction services stakeholder (baseline, affluent West) suggested that professionals who use alcohol as a ‘de-stressor’ may turn to ‘something prescribed’, with GPs being ‘put under more pressure to prescribe something’ by patients reporting stress. An addiction services stakeholder (baseline, deprived East), stated that ‘you could see an increase in things like XANAX^® [a benzodiazepine used recreationally] [alprazolam; Xanax, Upjohn UK Ltd, Kent, UK]’. One GP (baseline, deprived East) described alcohol and drugs use as ‘quite fluid already’, explaining that ‘I’ve worked here for over 20 years and to me, in the last 5 years there has been much more fluidity around drug use than previously’ and highlighting the risk of prescription drugs, such as the antiepileptic medications gabapentin (Pfizer, New York, NY, USA) and pregabalin (Pfizer), interacting with alcohol. In addition to identifying a risk of increased benzodiazepine abuse, an addiction services stakeholder (baseline, deprived West) identified methylated spirits as a potential substitute.

Some participants were dismissive of the risk of MUP increasing drug abuse. A police stakeholder typified this:

No I don’t think there will be a shift of substances. If you’d asked me that two years ago, I might have said yes absolutely because legal highs was a massive problem for us but that’s kind of died off to a certain degree and I don’t see anybody switching from alcohol to legal highs or other drugs. Even in terms of the youth drinking scene.

Baseline, deprived East, police stakeholder

A GP expressed doubt that MUP would increase the relative affordability of explicit substances sufficiently to affect drug abuse:

Drugs have always been expensive compared to what people are looking at with alcohol. I’m not sure with the minimum pricing legislation it’s going to become sufficiently expensive.

Baseline, affluent West, GP

The same participant later stated that ‘I can’t imagine that someone decides to go from alcohol to heroin, just ’cause the alcohol gets a bit more expensive. It’s too big a leap’. Similarly, an addiction services stakeholder (baseline, affluent West) suggested that cannabis is ‘quite expensive’ and, therefore, unlikely to be used as a substitute for alcohol.

Observed outcomes

Despite some stakeholders expressing concerns that MUP would drive use of alternative sources of alcohol and alternative substances, no stakeholders reported observing such outcomes. One youth worker suggested that local, potentially unlicensed alcohol trade may be in decline:

Dial-a Booze [www.dialaboozeglasgow.co.uk/; Glasgow, UK] I don’t think is as popular, because I think the cost, they can get it for – it’s expensive enough. I think they’ll get what they’ve got now and once it’s done it’s done.

Follow-up, deprived West, youth worker

Similarly, a licensing authority stakeholder explained that:

There was a persistent rumour that there was [a white van] going around [area], selling cider cheap. But myself and the police have looked in to it, we’ve no seen [anything].

Follow-up, affluent West, licensing authority stakeholder

Regarding cross-border sales, a police stakeholder stated:

We haven’t heard of anything at all, which is kind of surprising.

Follow-up, deprived West, police stakeholder

This participant went on to explain that ‘We said maybe [people] would be going to England and filling up a van and selling it, but we have no knowledge of that’. Similarly, an addictions services stakeholder (follow-up, affluence West) reported having heard nothing about cross-border sales.

When asked if they had observed an increase in drug abuse, a primary care stakeholder replied:

Interestingly not actually. That is a very interesting point there, because you could argue that a bag of heroin at £10 is still £10, which is actually cheaper than a bottle of Frosty Jacks now. So yeah, I haven’t, but that could well happen. I see people on 4 to 8 weeks, it’s only now that I’m beginning to hear about changes in addiction behaviour.

Follow-up, deprived East, primary care stakeholder

Similarly, a GP had observed no change, but acknowledged the limitations of their personal observations:

I can’t say I’ve noticed any difference, but as I say I’ve not measured anything.

Follow-up, deprived West, GP

Anticipated outcomes

A minority of participants discussed different potential economic impacts of MUP. An addiction services stakeholder expressed concern that large retailers, and not the government, stand to profit from the measure:

It’s not duty that they are putting on. So, if Asda, for talking sake, are charging across the board for all the alcohol, then that is a massive increase in price for some drinks. so, does that make Asda richer?

Baseline, affluent West, addiction services stakeholder

Later, the same participant framed that perceived injustice in terms of ‘the unplanned care costs that people with alcohol problems cause the NHS’, suggesting that additional customer expenditure on alcohol should fund alcohol services instead of adding to retailers’ revenue. One GP (baseline, deprived East) expressed concern that ‘little shops . . . that people depend on, that are perhaps already under pressure cos of Tesco’s [Tesco plc, Welwyn Garden City, UK] and Lidl’s [Lidl Stiftung & Co. KG, Neckarsulm, Germany] . . . will go under if quite a few of the sales are due to alcohol’, whereas another GP (baseline, deprived West) suggested that MUP might ‘have an effect on, you know, for example, the Scottish whisky industry or off license sales’.

Licensing authority stakeholders from both the East and the West expressed concerns about independent shopkeepers being unprepared for MUP. One such stakeholder explained that these retailers ‘won’t have an idea of what’s coming’ and are ‘worried that they’re going to get caught out’. This stakeholder presented three examples of the pressures on underprepared shopkeepers. In the first example, a shopkeeper was distressed by being pressured by Trading Standards to adjust her price marking, in a manner that the stakeholder regarded as ‘undiplomatic’ and unlikely to ‘create good relations’. In the second example, the same stakeholder described shopkeepers needing to sell stock at discounted prices to clear it prior to the implementation of MUP:

It’s not enough time for them to get rid of their stock so they’re now sitting with stock that they’re having to try and sell off at a discount just now before the 1 May or come after the 1 May they’re not going to be able to get rid of it.

Baseline, affluent West, licensing authority stakeholder

In the third example, the stakeholder described retailers mistakenly buying discounted alcohol unaware that impending price rises may render the products unable to be sold:

We’ve had at least one retailer in this area hoovering up Frosty Jack cider because they’re all selling it off cheap without realising I can’t sell this after the 1 May. If I don’t get shot of it by the 1 May.

Baseline, affluent West, licensing authority stakeholder

Observed outcomes

A small number of participants perceived MUP as having been economically beneficial to independent retailers. An addictions services stakeholder reported having read that:

Your corner shop saw an increase in their alcohol sales because the supermarkets couldn’t do the big deals. So, they were actually – in terms of economically, they were benefiting. It didn’t pay people to go to the supermarket because they could almost offer the same price.

Follow-up, affluent West, addiction services stakeholder

This was reinforced by licensing authority stakeholders. One such stakeholder (follow-up, deprived East) described how ‘a shopkeeper said his sales have actually gone up because a four pack of whatever it was is now the same in his shop as it is in Asda’. Another licensing authority stakeholder (follow-up, deprived West) echoed this, explaining that ‘My understanding is it has slightly evened out the market. It’s slightly more favourable for the small retailers than the bigger retailers’.

Licensing authority stakeholders identified some temporary problems for independent retailers related to the implementation of MUP. One participant (follow-up, affluent West) reinforced concerns from the baseline interviews about unprepared shopkeepers suffering economically because of finding themselves with stock they are unable to sell:

So there was at least four or five of my premises had that problem and that’s what I had to tell them. I said, you will need to get rid of it before the 1st otherwise you are going to be stuck with it. You are going to have to try and sell it for £12 a bottle.

Follow-up, affluent West, licensing authority stakeholder

Similarly, a stakeholder in the deprived East study area experienced resistance from shopkeepers around the pricing of white ciders:

That was a problem that I came across a lot. A lot of resistance from the trade because they felt I was going to put them out of business. Well not me personally but I was the one dealing with it, because they had high sellers, the strong ciders and a bottle that was price marked £3.99 was suddenly going up to £11.25 and I had a lot of people shouting at me saying you’re putting me out of business.

Follow-up, deprived East, licensing authority

The same stakeholder identified that, although bottles may be correctly price marked, tills may be set to sell them more cheaply:

There is another dimension to this that no one has even thought about, is that whilst a bottle of vodka might be displayed at £13.13 it’s what gets rung through the till. I mean that’s the unknown. So, they may have a till pre-set because its maybe not quite kosher, so rather than £13.13 they’re getting it for a fiver. They’re drinking that and then there have been instances of people going blind.

Follow-up, deprived East, licensing authority

Perceived effectiveness

Many participants expressed doubts about the policy’s ability to tackle alcohol dependency, with many fearing that it may have potential adverse consequences for drinkers with dependency issues living in deprived circumstances. These narratives were dominated by predictions of the policy’s impact on crime and, in particular, on shop lifting:

They will just go to bigger lengths to get it, because if people have an addiction people rely on something they will always find a way to get it. So whether that’s stealing money or stealing the item or whatever it is, they will do it.

Wave 1, FG01 – women, 18–24 years, ABC1, West

If you’ve not got the money then you will need to find the money and what will that cause, will that not cause more crime because people will need to shop lift, they will need to do this, because people aren’t going to go, ‘Oh it’s that price I will need to stop then!’ . . . there will be a lot of desperate people running about.

Wave 1, FG08 – men, 30–55 years, DE, West

More minor themes included increases in street begging, sacrificing food purchases, consuming other products that contained alcohol [e.g. Brasso (Reckitt, Slough, UK), methylated spirits, hairspray] and getting into debt (e.g. spending the rent money). There were few mentions of violent crime (e.g. customers being abusive towards shopkeepers) or narratives describing wider social impacts on the likes of families or communities:

I do know that even families which are, which have alcohol problems will still want to subsidise it, . . . they may well already have social work care or not yet have social work care, [but] social work intervention will end up coming in because the money has got to come from somewhere. If you are an alcoholic you are going to get the money for that drink . . .

Wave 1, FG11 – women, 30–55 years, DE, East

It’s like we were saying, the people who drink that [white cider], want to drink it, they will find a way to spend £11 whether they miss food for a week, or they’re maybe stealing, or selling stuff, like if they really really want to buy it and drink it, they will find a way.

Wave 2, FG09 – women 18–24 years, DE, East

If you’ve got someone who’s homeless and they’ve got £5 there instead of spending £2 on that and then having £3 to get something to eat they then will sacrifice the food and just get what . . . that comes first, the alcohol comes first for them and then the food. So effectively in trying to help people it’s actually backfired a little bit on certain people.

Wave 2, FG09 – women, 18–24 years, DE, East

Substituting alcohol with drugs also emerged as a minor theme, but was mainly raised in connection with youth drinking, where drinking and drug-taking was reported to be relatively common:

The price of drugs aren’t changing so they [students] might end up doing that a bit more . . .

Wave 1, FG02 – men 18–24 years, ABC1, West

[When] you take cocaine the drink doesn’t hit you, whisky tastes like water, you’re guzzling it.

Wave 1, FG06 – men, 18–24 years, DE, West

The potential for cross-border shopping also emerged as a common theme, although there were few reports of this actually occurring post implementation, with distance thought to be a governing factor:

I’ve not heard of anybody doing the beer run yet, but I thought that would have materialised . . .

Wave 2, FG04 – men, 30–55 years, ABC1, West

I dare say it’s happening closer to the border I don’t know if people would justify driving all the way down there and back for the difference in the price.

Wave 2, FG08 – men, 30–55 years, DE, West

Participants from the two deprived study communities tended to take more extreme or emotive positions on the policy, in some cases rejecting it outright when considering possible adverse consequences for disadvantaged dependent drinkers:

How do you feel about this policy, this unit price thing?

It’s ridiculous . . . Especially for that Frosty Jack, that’s . . . That is just ridiculous . . . It’s outrageous . . . I actually feel sorry for the alcoholics that normally drink that because that would make them want to drink like!

Wave 2, FG11 – women, 30–55 years, DE, East

In contrast, those from the affluent study community were more likely to assume more nuanced positions, for example characterising the policy as something that was worth trying or that they’d want to see evidence of effectiveness before making any judgments:

It’s kind of targeted at poorer drinkers who are more likely to be problem drinkers. So I can understand why they’re doing it, to stop them falling into that trap of buying cheap drink. But the ones who can only afford cheap drink I worry for what . . . they can get black market stuff or try.

Wave 2, FG04 – men, 30–55 years, ABC1, West

I think they would have to justify it, you’d have to show me some statistic and prove that it’s making this a great deal. And I think people would be more inclined to get behind it . . .

Wave 2, FG04 – men, 30–55 years, ABC1, West

Participants from deprived study communities also tended to demonstrate greater empathy and identification with dependent drinkers living in deprived circumstances. In some instances, participants described what they saw as striking inequities in the way in which the policy targeted these vulnerable groups:

They are just trying to rip the poor off again aren’t they, guaranteed because we are not able to fight against it.

Wave 1, FG12 – men, 30–55 years, DE, East

See the people that is on the bread line they are going to be punished and all so they are when this comes out, because that’s their only outlet, they might just have put the wains to bed, single parent, wee bottle of wine, take that away, where is that guy going to get [his drink].

Wave 1, FG08 – men, 30–55 years, DE, West

Participants from all socioeconomic groups tended to believe that the policy was more likely to have a positive impact on underage drinking and were more likely to support it on these grounds:

I think it’s a good thing for like underagers, like to try and stop it, because of the Frosty Jacks, I think it’s good in that sense.

Wave 2, FG01 – women, 18–24 years, ABC1, West

However, as previously indicated, some highlighted the potential for youth drinkers to switch from alcohol to other mood-altering substances for recreational purposes:

See the likes of the younger ones, I think that they are going to get out of their face somehow or other, and . . . if they are not drinking it could be something else that they are doing that might be even worse. If they can’t afford the bottle of cider anymore then some drug dealer . . .

Wave 1, FG08 – men, 30–55 years, DE, West

In contrast, it was often believed that drinkers with alcohol dependency were more likely to adopt other strategies and measures, such as theft or sacrificing food purchases, to fund their alcohol habit, rather than switch to other less expensive substances. However, this view was not necessarily supported by those who had more direct experience of addiction:

That’s what happened to me . . . the likes of, you just move from one to the next, so if you canny afford that . . . I went away from alcohol to drugs and drugs to alcohol again eh . . . It’s the cheapest addiction you go for, what you can afford do you know what I mean?

Wave 1, FG12 – men, 30–55 years, DE, East

Perceived impact on prices

Support for the policy was also mediated by concerns and anxieties about how it would impact the price of those drinks they typically consumed and the amount they spent on alcohol:

So what does it do to a vodka in a bar, how much is that going to cost me? I don’t know what would it cost me!

Wave 1, FG03 – women, 30–55 years, ABC1, West

They are trying to stop alcoholism which it probably would to be fair . . . But as long as Kopparberg’s [Kopparberg Breweries, Kopparberg, Sweden] and vodka are all right [the products I drink are unaffected].

Wave 1, FG10 – men, 18–24 years, DE, East

For our age group if we are not drinking on a regular basis then I wouldn’t, if it’s going to go up then I don’t mind paying it for one night during the week I don’t mind paying a bit more for a bottle of vodka because I know it’s only for that 1 day.

Wave 1, FG01 – women, 18–24 years, ABC1, West

However, some participants from deprived study communities appeared to find the prospect of MUP more financially challenging:

I am not an alcoholic but . . . I’m still going to buy it, whether it’s a certain price or not . . . It’s not going to stop me from getting what I like at the weekend.

Wave 1, FG05 – women, 18–24 years, DE, West

We will keep drinking some way or another we will find a way around about it or we will have the cash maybe to pay a wee bit extra or whatever . . .

Wave 1, FG08 – men, 30–55 years, DE, West

At the post-implementation follow-up discussions, most reports of any changes to the prices they were paying for alcohol were relatively small. However, many continued to voice concerns about potential unintended consequences for more vulnerable drinkers. These are examined further in the next section, perceived impact of the policy.

Stakeholder interviews

We recruited a baseline sample of 25 participants pre-policy implementation (in February 2018) and re-interviewed 16 of these participants after MUP had been implemented (in October 2018). We recruited five additional interviewees at follow-up to ensure that our sample included the diversity sought, meaning that a total of 30 professional stakeholders participated in our study. Most interviews were individual one-to-one interviews (n = 23), but a minority preferred paired or small group interviews. Our sample achieved diversity in relation to professional role, including addiction services, youth workers, primary care, police and licensing authorities. We achieved geographic coverage across our three areas of interest (two deprived areas and one affluent).

Overall, our qualitative interviews (which reported professional ‘intelligence’ from across Scotland, as well as local perspectives) suggested high compliance with the policy and that the policy was implemented as intended. Licensing authority and police participants concerns reported that retailers’ compliance with MUP had predominantly been good, with any issues of non-adherence to the policy being very temporary. Prior to the policy’s implementation, expectations of the impact on consumption were mixed, with some stakeholders anticipating long-term reductions in consumption, but others expressing concern that consumption by wealthy individuals and those with ‘chaotic lifestyles’ would not change. Prior to the policy, stakeholders expressed concern about the potential for a range of negative predominantly short-term outcomes, particularly in relation to the potential for alcohol withdrawal among dependent, low-income drinkers. However, these anticipated negative outcomes were predominantly not evident in stakeholders’ experiences subsequent to implementation. Similarly, prior to the policy’s introduction, there had been concerns of increases in criminality (e.g. smuggling) and adverse economic impacts. However, interviewees felt that these did not materialise and, indeed, felt that economic impacts were often positive, with the policy helping to create a level playing field for small shops struggling to compete against large supermarkets. Although experiences of MUP were largely positive, some potential negative unintended consequences were identified. For example, there were concerns that the increased profitability of selling alcohol could cause the re-emergence of a problem of overprovision of alcohol within some communities. However, most of these concerns had not been observed to have occurred, but were rather issues that may emerge in the future.

Young people and high-risk drinkers’ experiences

To develop an understanding of the experiences of high-risk drinkers, we collected data from qualitative focus groups because they provide an opportunity for participants to explore their views in free-flowing discussions, generating data through participant interaction and group dynamics. A total of 24 focus groups were conducted with young binge drinkers (aged 18–24 years) and older heavy drinkers (aged 30–55 years) living in three different communities in Central Scotland (two deprived areas and one affluent area). These drinkers were selected to represent key subgroups of the Scottish population who may be particularly affected by MUP policy and are of substantial policy interest. A total of 105 people took part, with 68 taking part in both sets of focus groups (i.e. pre and post implementation in March/April and October/November 2018, respectively). Eighty-four people took part at baseline, of whom 47 took part in repeated focus groups, with an additional 21 recruited to maintain diversity in participant characteristics at follow-up focus groups.

In general, participants reported that they did not notice any differences in prices before and after the policy was introduced. Our results suggest that MUP was understood to be particularly targeted at the prices of cheap alcohol, especially high-strength ciders and the groups of drinkers typically associated with these products, such as underage street drinkers and poor dependent drinkers. It was generally perceived that the policy was not targeted at the focus group participants. However, their appeared to be a disconnect between this view and understanding of the practicalities of how the policy operates, with many people believing that drinkers would simply switch to buying something cheaper (which is prevented by MUP). In general, support for the policy was facilitated by the belief that it was targeted, especially at underage drinking, although there were concerns that dependent drinkers would maintain their consumption through criminality or reducing other essentials (e.g. food). Following implementation, evidence of such adverse consequences was largely absent, including in the more deprived study communities that had expressed greater concerns previously.

Primary outcome: drug consumption

We have shown that there is no evidence that, overall, illicit drug consumption changed in this population as a result of MUP. In this component, the general trend was for no overall increases in illicit drug use, although there was an increase in cannabis use. Routine data shows that cannabis use in the last 30 days had fallen for 15-year-olds from 13% in 2006 to 8% in early 2018 in Scotland,171 whereas in England the proportion of young people who reported having ever used cannabis in their lifetime decreased from 41% in 2002 to 21% in 2018, although since 2014 the proportion among boys increased slightly. 172 Therefore, the pre-MUP trend was opposite to the trend we found with MUP, indicating a stronger effect of increasing cannabis use from MUP than if cannabis use had been increasing anyway.

Secondary outcomes: alcohol consumption

The greater odds of any alcohol purchase in the last month from on-licensed premises were not unexpected by the authors, as MUP almost exclusively affected off-license sales of cheap high-strength alcoholic beverages, and not the more expensive on-license trade. The Sheffield alcohol policy model82 predicted a fall in tax revenue of £15M, with £12M attributable to the off-trade and £4M to the on-trade (rounded figures). Although this was not a move to on-license purchase, it does predict a smaller effect on the on-trade.

The possible effects in substrata have implications for health inequality, but the only substantive change was the greater odds of those aged < 19 years purchasing alcohol from both on- and off-license sources. MUP is likely to have an impact on this group, as the price increase was not seen by them as a barrier to purchase and, furthermore, most products preferred by under 18s, who were part of our sample, were not affected by MUP. 100

O’Donnell et al. 88 recently reported an interrupted time series analysis of alcohol off-trade purchases data, comparing England and Scotland before and 8 months after MUP implementation. The study showed a clear immediate reduction in alcohol purchases in Scotland, compared with England, after MUP, but some uncertainty as to whether or not the effect was sustained. Purchase and sales data also give no indication of consumption stratified by social group. Owing to the different study populations, ours being younger and better educated and with more females, our results do not contradict those of O’Donnell et al. 88

As MUP is likely to have a disproportionate effect in reducing consumption of the low-cost, high-strength alcoholic drinks favoured by high-risk alcohol consumers,82 the policy is focused on those most at risk. Public health evidence about the most effective ways of reducing health inequalities64 intimated that, as a structural approach that does not require individuals to take up an intervention, MUP would also be an effective way to reduce health inequalities.

In conclusion, overall, we found no drug substitution effect arising from MUP, although there was a greater likelihood of alcohol misuse. There were some effects within substrata by age and a greater likelihood overall to have purchased alcohol from an on-licensed premises in the past month. There may be implications for young people’s health. Our evaluation,39 which itself contains multiple components, is part of a wider programme co-ordinated by Public Health Scotland and should be understood in the context of those results. 173

Group composition

A presentation of our results provided the basis for an interactive discussion of the validity of our findings for each study component. There were five participants who were recovered alcohol users, three females and two males, aged between 20 and 70 years, and one worked for a non-statutory addiction support service. Participants appeared not to be from an affluent area. The session was in depth, lasting nearly 2 hours. The point was made that longer-term chronic effects such as liver cirrhosis would not be captured in our study and we cannot report anything about long-term results.

The key points arising were under three large themes: (1) the lack of effects, especially on dependent drinkers, (2) suggestions for adjustments to the approach and (3) views about whether or not the MUP should be increased.

Lack of effects

Participants were not surprised that there was little evidence of any effect from MUP, although there were anecdotes of changing beverage from white cider to spirits and people buying maybe one can less per week (which was still a reduction). As a positive, one participant suggested that people might be more likely to ask for help if they could not afford to drink as much as they needed. There was some surprise about the increased use of pubs and other on-licensed sources of alcohol. There was no surprise about the lack of effect on illicit drug consumption also, although it was agreed that this could be seen as positive. One participant cited the relative cheapness of a bag of heroin compared with a bottle of vodka, and participants agreed that more was drunk when combining drugs and spirits.

Two respondents (both females) who did not drink white cider saw no effects on their personal consumption from MUP.

There was a consensus that price would not be a deterrent for people who were chronically dependent on alcohol. Long-term drinkers would keep up their consumption and cut spending on food and fuel. However, it was thought unlikely that increased food bank use was due to MUP. It was also felt that small increases in price would also fail to deter weekend drinkers.

Adjustments to minimum unit pricing

It was suggested that beer and cider containing > 5.5% of alcohol should be banned. This would complement strength caps that already exist for wines and spirits, and would affect some of the beverages (e.g. alcopops) favoured more by young people, which participants agreed should be harder to buy.

More education about the dangers of alcohol was suggested for school children.

Alcohol was thought to be overpromoted and glamourised and to be ingrained in the Scottish culture, and this needed to be countered and promotion more restricted.

Supermarket sales of alcohol should not be allowed and licensing should be more restricted [Queensland (Australia) and Canada were mentioned as examples of alcohol being sold only from designated outlets, and pharmacies and garages mentioned as inappropriate places to sell alcohol].

The minimum unit pricing level

When asked if the minimum price should be raised from £0.50 per unit, the consensus was that people did not know, but probably not.

There was a generally high level of concern about what happened to the extra revenue. The view was that this should go not to retailers, but to services and support for dependent drinkers.

Finally, although MUP or some form of price intervention was seen as good for young people, it was thought to be only a small part of the solution to the alcohol problem.

Search strategy

1. (Alcohol and minimum and unit and price).mp. [mp = ti, ab, ot, nm, hw, fx, kf, ox, px, rx, ui, an, sy, tc, id, tm, mh] (88)

2. limit 1 to abstracts (76)

3. limit 2 to full text (15)

4. limit 3 to yr = “2012 -Current” (14)

5. remove duplicates from 4 (12)

Search strategy

1. (Alcohol and minimum and unit and price).mp. [mp = ti, ab, ot, nm, hw, fx, kf, ox, px, rx, ui, an, sy, tc, id, tm, mh] (88)

2. limit 1 to abstracts (76)

3. limit 2 to full text (15)

4. limit 3 to yr = “2012 -Current” (14)

5. remove duplicates from 4 (12)