The use of existing data sources to evaluate the impact of alcohol control policies or contextual factors affecting alcohol consumption in the UK and in Sri Lanka

Nugawela Pathirannehelage, Manjula Darshani (2017) The use of existing data sources to evaluate the impact of alcohol control policies or contextual factors affecting alcohol consumption in the UK and in Sri Lanka. PhD thesis, University of Nottingham.

[img] PDF (Thesis - as examined) - Repository staff only until 31 December 2019. Subsequently available to Repository staff only - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Download (7MB)

Abstract

Background

Harmful use of alcohol is the fifth leading risk factor for global burden of disease, disability and death. While the level of alcohol consumption varies around the world, it causes numerous preventable health and social issues in many countries and around 3.3 million deaths per year globally. In the UK alcohol consumption is decreasing; by contrast, alcohol consumption in Sri Lanka has been rapidly increasing over recent years. However, in both settings alcohol misuse represents a major public health concern. Therefore, it is important to evaluate existing alcohol control policies or contextual factors affecting alcohol consumption in these settings and to identify existing data sources that can be used for alcohol control policy evaluation.

This thesis aimed to investigate the suitability of existing data sources in the UK, identify potentially suitable measures and use those measures to evaluate the impact of Licensing Act 2003 in England, which allowed flexible opening hours including 24-hour drinking at on-trade premises in England and Wales from November 2005 onwards. This thesis further aimed to apply the lessons learned from the UK to Sri Lankan context, identifying potential data sources and using these to evaluate the effect the end of conflict in 2009 on alcohol consumption in Sri Lanka.

Methods

A range of sources of data on alcohol consumption and consequences in the UK were reviewed to identify those appropriate for alcohol control policy evaluation, and in particular for time series analysis which requires consistent data collected at regular intervals for a long period of time. The suitability of UK primary care data on alcohol consumption was assessed by identifying the proportion of patients with a record of alcohol consumption status in the last year, as well as ever since their registration with a practice. The quality of alcohol consumption measures collected by English national surveys was assessed by comparing them with the international guidelines for measuring alcohol consumption in population surveys. Existing data sources on alcohol consumption and consequence in Sri Lanka were also identified and reviewed to identify their suitability for alcohol control policy evaluation. Interrupted time series analysis was then conducted on these measures to establish the effect of the Licensing Act 2003 on alcohol consumption among adults (age≥16) in England, and the end of the conflict on recorded alcohol consumption among adults (age≥15) living in the areas that were not directly affected by the conflict in Sri Lanka.

Findings

In the UK, there are numerous sources of alcohol consumption and consequence data. However, few provide frequently collected data from large samples over long time periods for time series analysis. The recording of alcohol consumption in primary care remains low, particularly when recent recording within a given year is considered. Moreover, alcohol consumption recording in primary care is higher among at-risk groups such as women in child bearing age, older men and women who are likely to have an illness linked to alcohol. Therefore, primary care data are currently unsuitable for alcohol control policy evaluation. Comparison of the alcohol data collected in English national surveys with recommendations from international guidelines showed that they have failed to maintain the consistency of data collection over time and to collect information on some of the key alcohol consumption measures such as the frequency of binge drinking. However, Health Survey for England (HSE) measured alcohol consumption on the heaviest drinking day of the last week consistently over time and this measure was available on a quarterly basis for a nationally representative sample.

Interrupted time series analysis of HSE heaviest drinking day data from 2001 to 2013 showed that prior to the implementation of the Act alcohol consumption among adult male and female drinkers remained constant at around 8.3 units and 5.5 units of alcohol respectively. After the Act, there has been a gradual decline (less than 0.03 units per quarter) in the heaviest drinking day consumption among both male and female drinkers in England. However, it is difficult to attribute this decline in consumption to the Licensing Act as there was no step change in the consumption soon after the Act and the decline in heaviest drinking day consumption was small and gradual over a period of seven years.

When compared with the UK, Sri Lanka has a limited number of data sources on alcohol consumption and alcohol consequences. The Department of Excise alcohol sales database was identified as the only data source that can be used for alcohol control policy evaluation in Sri Lanka. However, these population-level data cannot be used to identify the characteristics of people who drink and patterns of drinking such as binge drinking.

Interrupted time series analysis showed that recorded alcohol consumption among Sri Lankans living in areas that were not directly affected by the armed conflict increased markedly after the end of the conflict in 2009, with a dramatic acceleration in the trend of adult per capita consumption. Annual adult per capita alcohol consumption among Sri Lankans increased from 1.59 litres of pure alcohol in 1998 to 2.56 litres of pure alcohol in 2013. Prior to the end of war in 2009 adult per capita alcohol consumption was increasing by 0.051 litres of pure alcohol per year (95% CI 0.029-0.074, p<0.001). After 2009, it increased by 0.154 litres per year (95% CI 0.082-0.226, p=0.001). it increased by 0.166 litres of pure alcohol per year (95% CI 0.095-0.236, p<0.001), almost a three-fold increment in the increase per year compared to the trend prior to the end of the conflict. Beer consumption showed the highest per capita growth compared with other beverages.

Conclusions

This thesis identified the existing data sources that can be used for alcohol control policy evaluation purposes in two settings; UK and Sri Lanka. It has highlighted the further improvements required in existing alcohol consumption related data sources in both countries and discussed the potential of applying lessons learned from the UK context to Sri Lankan context. Despite the current trend in alcohol consumption, both countries experience a significant public health burden due to alcohol misuse. Therefore, both countries will require formulation and implementation of new policy measures. However, Sri Lanka does not have high-quality individual level alcohol consumption data to support the monitoring and evaluation of alcohol control policies. Therefore, this thesis has emphasised the need to generate high-quality alcohol consumption data in Sri Lanka and carry out monitoring and evaluation of alcohol control policies to tackle the alcohol-related burden.

Item Type: Thesis (University of Nottingham only) (PhD)
Supervisors: Langley, Tessa
Szatkowski, Lisa
Lewis, Sarah
Keywords: Alcohol Consumption, Data sources, UK, Sri Lanka, Surveys, Primary Care Data, Conflict Exposure
Subjects: W Medicine and related subjects (NLM Classification) > WA Public health
W Medicine and related subjects (NLM Classification) > WM Psychiatry
Faculties/Schools: UK Campuses > Faculty of Medicine and Health Sciences > School of Medicine
Item ID: 44806
Depositing User: Nugawela Pathirannehelage, Manjula
Date Deposited: 15 Dec 2017 04:40
Last Modified: 16 Dec 2017 02:10
URI: http://eprints.nottingham.ac.uk/id/eprint/44806

Actions (Archive Staff Only)

Edit View Edit View