Hazardous, harmful and dependent drinking among adults living in England

This is an extension request to the Data Sharing Agreement until 22nd June 2025 to enable continued access to APMS 2014 data to conduct a more detailed analyses of the alcohol-related data following the release of the 2014 APMS Main Report in 2016. The researcher was a lead author on the Chapter 9 report (Drummond, McBride et al. 2016 in McManus et al. 2016). This work is part of ongoing academic research and is not externally funded.

The purpose of the project is to further the understanding of hazardous, harmful and dependent drinking among adults living in England. Annual per capita alcohol consumption in England (ie. the number of litres of pure alcohol consumed by each individual aged 15 years and older) has fluctuated over recent decades, increasing from 9 litres in 1975, to a peak of 11.5 litres in 2004, before falling to 10 litres in 2011. A number of drinking-related trends are hidden beneath this national picture of consumption, including: (a) larger proportions of the population now abstain from alcohol (particularly younger adults); (b) younger adults who do drink are drinking less; (c) women of all ages are drinking more; and (d) middle-aged and older drinkers are drinking more, particularly men.

In addition to these age-group trends, research also indicates a number of key differences between the drinking patterns of adults of different social class: people in lower social classes tend to drink less overall, but have more infrequent heavy drinking occasions, than those in higher social classes. Despite drinking less, people in lower social classes experience more alcohol-related harm (ARH), associated with their drinking, compared to those of higher social class. ARH refers to negative health and social consequences related to drinking, including use of health services (in particular hospital admissions and Accident & Emergency (A&E) attendances), health problems (e.g. alcohol-related cancer), relationship difficulties, inability to perform work or other duties properly, legal problems, and even premature death.

It is not clear why people of lower social class are at considerable disadvantage; one theory referred to as ‘multiple health challenges’ proposes that drinkers from lower social classes experience higher levels of ARH because they also engage in other negative health behaviours (e.g. smoking, poor diet, lack of physical activity, and obesity) or experience other challenges (e.g. mental health difficulties) that greatly increase their risk of experiencing poorer outcomes as a result of their drinking. This theory may help explain why, despite overall falling levels of alcohol consumption, the costs of treating ARH in the NHS are spiralling, from £2.7 billion/year in 2006/7 to £3.5 billion/year in 2014-15.

The lawful basis for processing data under GDPR has been reviewed against the guidance provided by IGARD and has been assessed as acceptable. The details are:

Article 6(1)(e) – legitimate interest under “public task”.

(processing is necessary for the performance of a task in the public interest or in the exercise of official authority vested in the controller).

University of Ulster is a public authority. University of Ulster's Royal Charter includes the following statement "The objects of the University shall be to advance education through a variety of patterns, levels and modes of study and by a diversity of means by encouraging and developing learning and creativity, for the benefit of the community in Northern Ireland and elsewhere; to preserve, advance and disseminate knowledge and culture through teaching, scholarship and research, and to make available the results of such research; and to promote wisdom and understanding by the example and influence of corporate life.

The University shall be a teaching, research and examining body and, subject to the provisions of this Our Charter and Statutes, and, insofar as is compatible with its status as a charitable body, shall have the necessary powers to further these objects, namely:

(A) To provide full-time and part-time instruction, leading to awards and qualifications or otherwise; to make provision for research and research training and for the preservation and advancement of knowledge in such manner and through such media as the University may determine."

‘Necessity’: Throughout the application process, the necessity of the processing for the performance of the task has been assessed. This included but was not limited to ensuring appropriate minimisation of the data to ensure that only the minimum amount of data required are processed. During the application process it has been considered whether the information that the processing aims to determine is already available from other sources or whether the task could be performed using publicly available data or data from alternative sources than NHS Digital. Consideration has been given to whether the volume of data being requested is proportionate to the expected benefit and, through examination of the expected benefits consideration has been given to whether the task is itself necessary. The applicants have been involved in this process and participated in teleconference with our data production team to agree which data products and fields were appropriate for use for the purpose of processing. Therefore, we are satisfied that this request is appropriate, necessary and proportionate for the performance of the task described in the Purpose statement and that there is no other reasonable means for the data processor to achieve their purpose that is less intrusive to the data subjects.

Processing data for the purposes of research is considered to be one of University of Ulster’s public tasks. The processing of the APMS dataset is considered necessary as there are no other means of examining the objectives in a less restrictive way. Individuals will not be harmed through the processing of the data.

Article 6(1) of the GDPR (lawfulness of processing), the reference in point (e) to processing of personal data that is necessary for the performance of a task carried out in the public interest or in the exercise of the controller’s official authority includes processing of personal data that is necessary for… (d) the exercise of a function of the Crown, a Minister of the Crown or a government department”. Universities which have a royal charter granting them a right to perform research would therefore be considered to be exercising a function of the Crown in undertaking that research.

Article 9(2)(j) – 'processing is necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes in accordance with Article 89(1) based on Union or Member State law which shall be proportionate to the aim pursued, respect the essence of the right to data protection and provide for suitable and specific measures to safeguard the fundamental rights and the interests of the data subject'.

Research is now urgently needed to provide a robust testing of this theory to determine whether patterns of alcohol use are interacting with important lifestyle factors, differently for adults of different age-ranges, sex, and social class, and who live and receive health services across England. The proposed research aims to achieve this by analysing data from three national mental health surveys – the Adult Psychiatric Morbidity Survey (APMS) - conducted in 2000, 2007 and 2014, with adults aged 16 years and older living in private households across England. Analysing this survey data with appropriate statistical models will provide a unique opportunity to thoroughly test the theory of multiple health challenges. It is expected that these research findings will be used in the future by the NHS to help (1) tailor healthcare resources to target specific groups of drinkers deemed to be at greatest risk of ARH living in different regions of the country; and 2) estimate the type and amount of resources required to reduce the costs associated with ARH, with a particular focus on inpatient hospital admissions, across the country.

The only organisations involved in this project will be academic institutions. Collaborating institutions for this programme of non-funded research include: King's College London and the National Centre for Social Research. Academics from King’s College London will not access or process the data. Four academics from these institutions were lead authors on the main Alcohol Dependence chapter (Chapter 10) in the 2014 APMS report. Other authors from the APMS investigating team and co-authors on the main APMS 2014 report will also likely collaborate on this programme of work. These authors will collaborate on authoring the two proposed research papers/outputs for this programme of work. Specifically, they will contribute to: (1) developing the rationale and background literature review for the paper; and (2) assist with the interpretation of the results derived from the statistical analysis, and the implications of same, for the papers. Ulster University will be the sole data controller who also process data for this programme of research. The research will not include other organisations such as Clinical Commissioning Groups or local authorities.

This is a standalone academic research project in the area of alcohol-related harm. It will not feed into a bigger programme of research.

Permission to access and analyse the 2014 APMS data from NHS Digital is essential to the conduct of this academic research. 2014 APMS survey is unique in its depth of assessment of hazardous and harmful drinking patterns and alcohol dependency, engagement in negative health behaviours (e.g. smoking, drug use), stress and diagnosable mental health conditions, comorbid physical health problems, health care service use, as well as measuring multiple aspects of socio-economic status (SES), including occupation status, educational attainment, deprivation, personal debt, and household composition. APMS is distinctive in its collection of alcohol use data via self-completion, which improves the validity of consumption data when compared using computer-assisted personal interviewing. It is the optimal data source to conduct this programme of research.

No elements of this research project will take place outside of the UK.

There will be no transfer of data between institutions or organisations. All data will be processed at Ulster University.

Only substantive employees of Ulster University will access and process the data. The researcher will log in securely to the UK Data Service to access and download the APMS 2014 dataset and associated documents once the necessary permissions have been granted. The APMS 2014 data will not be accessed or processed outside of the UK.

The data NHS Digital supplied (i.e. the 2014 APMS data) will be analysed carefully, as follows:

1. The original 2014 APMS dataset received from NHS Digital via the UK Data Service will be securely archived on Ulster University's internal data servers. Weekly back-ups of the original file will occur.

2. Only the variables required for the analysis in this programme of work will be copied from the original 2014 APMS dataset into a new Statistical Package Social Sciences (SPSS) dataset, which will be labelled and saved as a working dataset on an encrypted pen drive. This dataset will be used to prepare all variables for statistical analyses (e.g. cleaning of missing data; re-categorisation of existing variables) and for the conduct of statistical analysis. Only one version of this file will be kept; redundant or obsolete versions of this file will be destroyed.

The variables listed below will be selected from the original archived 2014 APMS dataset for the purposes of analysis.

Outcome variable - AUDIT-SCORE. This variable will be re-categorised as a four-category variable as follows:

1.No/Low risk will be defined as AUDIT scores 0-7

2.Hazardous alcohol use will be defined as AUDIT scores 8-15

3.Harmful drinking/mild or probable dependence will be defined as AUDIT score 16-20

4. Dependence will be defined as AUDIT scores 20+.

The following variables will be selected from the original archived 2014 APMS dataset as explanatory/predictor variables for the respondent:

Age in year bands (16-24; 25-34; 35-44; 45-54; 55-64; 65-74; 75+)

Sex (male vs. female)

Ethnicity (White; Black; Asian; Mixed)

Employment status (Employed; unemployed; economically inactive)

Household composition (1 adult, 16–59 no child; 2 adults 16–59 no child; Large adult household; 2 adults one or both 60+ no child; 1 adult 60+ no child)

Common mental disorders (CMD) score (count)

Smoking status (smoker in the past year; yes vs. no and number of cigarettes smoked on average a week)

Drug use (drug user in the past year, yes vs. no and use of drug types in the past year)

The 2014 survey will not be linked to any other data source; however, it will be compared - but not linked – to the 2000 and 2007 APMS surveys for the purposes of population-level trend analyses. Propensity score analysis will be used to equate the 2014 APMS survey to the 2000 and 2007 APMS on key demographic characteristics namely age, sex, and socio-economic status. Weighted cross-tabulations will be estimated to determine the prevalence of alcohol use and hazardous, harmful and dependent drinking in the total sample and in subgroups. These data will be used to test differences in prevalence between the surveys using 2-sided t tests for independent samples (p < 0.05 will indicate significant differences in the estimates between surveys). Regression models will test the extent to which changes in alcohol use and hazardous, harmful and dependent drinking during 2000-2014 can be explained by period effects (i.e. economic recession), cohort effects, age effects, and the impact of other individual-level factors (i.e. being of lower socio-economic status, having co-occurring mental health problems, smoking and other drug use).

For this programme of research, the 2014 APMS data NHS Digital supplies will be analysed in conjunction with the 2000 and 2007 APMS survey datasets, which are available to bona fide researchers from the UK Data service (https://www.ukdataservice.ac.uk/). No data linkage between the 2000 and 2007 APMS surveys to the 2014 APMS will be conducted. Comparative analysis will be conducted between the three APMS surveys to examine trends in alcohol use and associated harm across this 14-year period. Propensity score analysis will be used to equate the APMS surveys on key demographic characteristics (i.e. age, sex, and socio-economic status). Data involved in this programme of research - that is the three APMS survey series conducted in 2000, 2007, 2014 - are required to examine trends in alcohol use over time. As detailed above, the APMS survey series used comparable measurements of alcohol use and associated harm, which neatly facilitates exploration of trends in the prevalence and correlates of these health behaviours at a population level over time.

No data will be linked to record individual-level data requested from NHS Digital (in this instance, the 2014 APMS survey data). There will be no requirement and no attempt to re-identify individuals.

All organisations party to this agreement must comply with the Data Sharing Framework Contract requirements, including those regarding the use (and purposes of that use) by “Personnel” (as defined within the Data Sharing Framework Contract ie: employees, agents and contractors of the Data Recipient who may have access to that data).

There will be no data linkage undertaken with NHS Digital data provided under this agreement that is not already noted in the agreement.

Data will only be accessed and processed by substantive employees of the University of Ulster and will not be accessed or processed by any other third parties not mentioned in this agreement.

The findings of this programme of research will be written in research report form, and the following papers will be submitted for publication to two academic journals:

1. "Prevalence of hazardous, harmful or dependent drinking in England, 2000, 2007, and 2014: results from the Adult Psychiatric Morbidity Survey". (to be submitted to JAMA Psychiatry in 2022).

2. "Testing the validity of the alcohol harm paradox in three national surveys: findings from the Alcohol Psychiatric Morbidity Survey" (to be submitted to Addiction in 2022).

No other publications, outputs, presentations, or dissemination events are planned for this programme of research.

Publications emerging from this research will be available to key stakeholders responsible for alcohol policy and health service planning.

These two publications are intended for an academic audience and will be accessible to those who have access to the journal subscriptions.

APMS low numbers and suppression:

In order to protect patient confidentiality in publications resulting from analysis of APMS data users must:

· guarantee that any outputs made available to anyone other than those with whom this agreement is made, will meet required standards, including the guarantee, methods and standards contained in the Code of Practice for Official Statistics (http://www.statisticsauthority.gov.uk/assessment/code-of-practice/index.html) and the ONS Statistical Disclosure Control (https://gss.civilservice.gov.uk/statistics/methodology-2/statistical-disclosure-control/) for tables produced from surveys;

· apply methods and standards specified in the Microdata Handling and Security Guide to Good Practice (http://www.data-archive.ac.uk/media/132701/UKDA171-SS-MicrodataHandling.pdf) for disclosure control for statistical outputs.

All outputs will contain only data that is aggregated with small numbers suppressed in line with the HES Analysis Guide

The findings from this study, which are expected to be published in academic report format in 2022-2023, should contribute to updating the research evidence base on trends in alcohol use in the English adult (16 years and over, general population), which can be used to support health and social care efforts to tailor services to target specific groups of drinkers across England deemed to be of greatest risk of alcohol-related harm into the future.

The purpose of the 2014 APMS survey (generally) is to help inform and improve local and national planning for health and support services. The survey series has had a huge impact on our understanding of mental illness, substance dependence and suicidal behaviour, and their causes and consequences. APMS datasets are the only national source of information on rates of untreated mental illness. Advancing on the descriptive analyses published in the 2014 APMS Survey report (McManus et al. 2016), this programme of academic research will provide a detailed evidence as to current levels of harmful, hazardous and dependent drinking in the English adult population aged 16 years and over. It is expected also to provide identify levels of treatment seeking behaviour (both for medication and psychological therapy or counselling) for alcohol-related problems in the general population in 2014. These findings, which are expected to be generated through the analysis of nationally-representative household survey data, should be a robust evidence source for forward planning by policy makers and health care professionals tasked with reducing alcohol-related harm in England.

This research might not directly affect patients, although estimates from the NHS suggest that 9% of men and 3% of women experience alcohol dependence. The findings from this study, which are expected to be published in academic report format in 2022-2023, should contribute to updating the research evidence base on trends in alcohol use in the English adult (16 years and over, general population), which can be used by to support health and social care efforts to tailor services to target specific groups of drinkers across England deemed to be of greatest risk of alcohol-related harm into the future.

The impact of the analysis will be possible after the publication of research outputs emerging from this body of research circa 2022-23.

The significant delays to this project that have been encountered due to Covid-19 have, to date, precluded the expected benefits from being achieved. As it is anticipated that the expected benefits will be yielded following dissemination of the findings, no changes are being made to the expected benefits specified in the original application. However, as detailed in the updated ‘outputs’ section, the completion of analysis and therefore dissemination of findings has been significantly delayed. Therefore, the expected benefits will not be yielded until a later date. Updated target dates have been outlined in the updated ‘outputs’ section.