Critique 235: Does type of alcohol or drinking pattern explain greater adverse effects of alcohol consumption among lower socio-economic subjects? 15 January 2020
Gartner A, Trefan L, Moore S, Akbari A, Paranjothy S, Farewell D. Drinking beer, wine or spirits – does it matter for inequalities in alcohol-related hospital admission? A record-linked longitudinal study in Wales. BMC Public Health 2019;19:1651. https://doi.org/10.1186/s12889-019-8015-3
Background: Alcohol-related harm has been found to be higher in disadvantaged groups, despite similar alcohol consumption to advantaged groups. This is known as the alcohol harm paradox. Beverage type is reportedly socioeconomically patterned but has not been included in longitudinal studies investigating record-linked alcohol consumption and harm.
Methods: We aimed to investigate whether and to what extent consumption by beverage type, BMI, smoking and other factors explain inequalities in alcohol-related harm. 11,038 respondents to the Welsh Health Survey answered questions on their health and lifestyle. Responses were record-linked to wholly attributable alcohol-related hospital admissions (ARHA) eight years before the survey month and until the end of 2016 within the Secure Anonymised Information Linkage (SAIL) Databank. We used survival analysis, specifically multi-level and multi-failure Cox mixed effects models, to calculate the hazard ratios of ARHA. In adjusted models we included the number of units consumed by beverage type and other factors, censoring for death or moving out of Wales.
Results: People living in more deprived areas had a higher risk of admission (HR 1.75; 95% CI 1.23–2.48) compared to less deprived. Adjustment for the number of units by type of alcohol consumed only reduced the risk of ARHA for more deprived areas by 4% (HR 1.72; 95% CI 1.21–2.44), whilst adding smoking and BMI reduced these inequalities by 35.7% (HR 1.48; 95% CI 1.01–2.17). These social patterns were similar for individual-level social class, employment, housing tenure and highest qualification. Inequalities were further reduced by including either health status (16.6%) or mental health condition (5%). Unit increases of spirits drunk were positively associated with increasing risk of ARHA (HR 1.06; 95% CI 1.01–1.12), higher than for other drink types.
Conclusions: Although consumption by beverage type was socioeconomically patterned, it did not help explain inequalities in alcohol-related harm. Smoking and BMI explained around a third of inequalities, but lower socioeconomic groups had a persistently higher risk of (multiple) ARHA. Comorbidities also explained a further proportion of inequalities and need further investigation, including the contribution of specific conditions. The increased harms from consumption of stronger alcoholic beverages may inform public health policy.
Many studies suggest that poorer, less-educated subjects, those with lower levels of socio-economic status (SES), have more severe adverse effects related to alcohol consumption than do subjects with higher SES. Among the hypotheses to explain this phenomenon, as enumerated in the present paper, are (1) different drinking patterns (more binge drinking, especially of spirits), (2) other poor health behaviors associated with lower SES (e.g., smoking, obesity), and (3) poorer response of lower-SES subjects to survey requests and under-reporting of current and life-time drinking. The present study from Wales collected survey data of geographic and individualized level of deprivation and of drinking habits and related them to hospital admissions for wholly attributable alcohol-related conditions (ARHA).
The authors report that subjects from more deprived areas and/or more deprivation from individual responses had higher risk of having one or more ARHAs than did subjects with less deprivation. However, the number of drinks/week reported explained little of such differences in the risk of ARHA, while associated habits/conditions (smoking, obesity, self-reported health status, and mental health conditions) explained larger amounts of the differences shown between more-deprived and less-deprived subjects. Overall, the greater risk of adverse effects from alcohol associated with deprivation remains largely unexplained.
Comments by specific Forum reviewers: Most reviewers considered that the authors had good assessments of geographic and individual levels of deprivation, and had appropriate means of ascertaining hospital admissions defined as “fully attributable to alcohol.” They were limited in their ability to answer many questions by lack of statistical power. The study was based on a population-based representative sample of subjects, but there were limited numbers for analysis (7,042 less-deprived and 3,996 more-deprived subjects). The authors had to group the more-deprived 40% and the less deprived 60% to get significant results, and thus they state “We are underestimating the extent of inequalities between the more extreme ends of the deprivation gradient.” Further, there were no data on characteristics of subjects responding and those not responding to the request for the initial surveys, and only one half of those surveyed agreed to link their data with hospitalization records, factors that limit how representative their results are for the public, even for their population.
Reviewer Ellison had questions about the fact that some subjects contributed data from multiple hospitalizations, and reviewer Thelle noted: “I agree that counting the same heavy alcohol user several times may induce biased results. The next visits could be collider effects.” However, reviewers Zhang, de Gaetano, and Janssen thought that the authors had dealt with this adequately in their analyses. Reviewer Zhang noted: “It seems to me that the authors tried to deal with the issue of using data from a subject more than once by using ‘a recurrent event model’ with admission as the outcome and using age as the underlying timescale rather than calendar time. If I were them, I would have considered the number of hospitalizations as a count data and analyze it using Poisson regression model (time of follow up as an offset). By doing so, they could also account for correlation issues within the same family (i.e., GEE with Poisson Link).” Stated reviewer Janssen: “The models they used as detailed in the methods were stratified by total number of admissions and took into account repeated admissions. That seems alright; they did not count people twice.” However, reviewer Thelle noted: “The repeated hospital admissions are reflecting both the social situation, comorbidities, as well as alcohol abuse.”
Inadequate assessment of alcohol consumption: A more serious concern related to the potential inadequacies in their assessment of alcohol intake. They based the frequency of drinking only on the “heaviest drinking day in the past week.” Further, they acknowledge that some (? many) of the subjects reporting no alcohol may have been “sick quitters.” Stated reviewer Zhang, “I agree that measurement error for alcohol consumption (frequency, pattern) is probable, and is likely to be larger than that of smoking and BMI.”
Forum member Thelle noted: “Using only the past week as exposure is probably underestimating the prevalence of binge drinkers. Also, the separation of former drinkers from life-time abstainers should be a sine qua non in today’s alcohol research.” Forum member de Gaetano and his colleague, Augusto Di Castelnuovo, wrote: “The most serious limitation of this study in our estimation is their lack of ability to separate former heavy drinkers from others not consuming alcohol within the past week; thus, the non-drinking group is a mixed category. We also express concern about using data from just ‘the heaviest drinking day in the past week;’ the title of the paper should have specified that it is dealing with heavy alcohol consumption.” Forum member Stockley added: “I also express concern about using data from just ‘the heaviest drinking day in the past week’. The ‘sick quitter’ is a critical missing component of the analysis, leading the authors to unsupported supposition.” Reviewer Janssen agreed: “They also make many assumptions, like proportional hazards. I agree it’s hard to verify those for the repeated events models, but they could have provided some sensitivity analyses, at least in an appendix.”
Effects of beer, wine, and spirits: As for their results for effects of type of beverage, it was noted that subjects with greater deprivation much more commonly consumed spirits but very little wine. However, the authors state in their discussion that “the type of beverage was not important over and above the number of units relating to inequalities.” It is unclear what the data are that support this statement, as in most results shown the hospital admissions were greater for spirits consumers than for beer or wine drinkers. It is interesting that spirits consumption occurred primarily in the youngest age group (aged 16-29). Further, it was stated that subjects with more deprivation consumed primarily beer and spirits, but little wine.
Effects of co-morbidities: Forum members agreed that co-morbidities, rather than just the number of alcohol units consumed, may explain much of what the authors call the “Alcohol Harm Paradox.” Ellison noted: “I am unsure how adequate the estimates of alcohol intake are for the frequency and pattern of drinking. (The authors state that alcohol frequency explained only 4% of the higher rate of ARHAs in the more-deprived group, but adding smoking and BMI explained 35.7%.). Reviewer Svilaas pointed out that the most important lifestyle related to hospital admissions was smoking: “People with a low SES are much more likely to smoke cigarettes.”
Forum member de Gaetano wrote: “Another limitation is the observation that comorbidities explained most of the inequalities and need further investigation. In any case, the analysis of single food items (such as alcohol) should be completed by the evaluation of full dietary behavior and the inclusion of clinical outcomes as complete as possible. This is why we always include all-cause mortality in our analyses.”
Forum member Mattivi wrote: “The results of this analysis are much in line with the conclusions of the study of Bellis et al, who found that ‘Deprived increased/higher drinkers are more likely than affluent counterparts to consume alcohol as part of a suite of health challenging behaviours including smoking, excess weight and poor diet/exercise.’ In other words, the potential harmful effects of alcohol cannot be considered independently from the overall life-style. Inequalities in food and diet have been investigated in several regions, including Wales, leading to the conclusion of Dowler that ‘Members of low-income households in the UK are more likely to have patterns of food and nutrient intakes that are less inclined to lead to good health outcomes in the short and long term.’”
Stated reviewer Van Velden, “This paper underlines that the effect of alcohol consumption must not be seen in isolation, but that we must take a holistic view in this regard. Low SES groups tend to have a poor diet with micronutrient deficiencies, as well as lower physical activity, that may contribute to the adverse effect of the associated alcohol-related harm. All these co-morbidities may account for the adverse health effects observed. It all boils down to a responsible lifestyle with moderation as the key to health maintenance.”
Forum member Skovenborg summarized two important implications of the results of this study, writing: “I agree with the comments given so far and have only a few points of interest to add:
- The phenomenon of “the alcohol harm paradox” is not restricted to alcohol; studies have found the phenomenon in obesity: mortality rates in non-smoking severely obese women were highest in the lowest occupational classes (Hart et al) and also regarding to the effects of smoking: for individuals living in the affluent Downtown neighbourhood, the relative odds of being unhealthy among smokers compared to non-smokers was less than one-half of the corresponding relative odds in the rest of the city (Birch et al). Data from the 1990 U.S. National Health Interview Survey also supported the vulnerability hypothesis of low socioeconomic status – that smoking inflicts greater harm among disadvantaged groups (Pampel & Rogers).
- Data from the longitudinal health and retirement study found an increased mortality risk of 2.84 (95% CI 2.35-3.60) in the most disadvantaged quartile of SES compared with the least disadvantaged quartile. Together smoking, alcohol consumption and physical inactivity explained 68% (35-104%) of this association leaving a risk ratio of 1.59 (1.03-2.45) for low SES (Nandi et al). In the Whitehall II study health behaviours (smoking, alcohol consumption, diet and physical activity) attenuated the association of SES with mortality by 75% (95% CI 44%-149%) but only by 19% (95% CI 13%-29%) in the French GAZEL Study (Stringhini et al). Unhealthy lifestyle behaviors are likely to be major contributors of socioeconomic differences in health only in contexts with a marked social characterization of health behaviors.”
References from Forum critique
Bellis M, Hughes K, Nicholls J, Sheron N, Gilmore I, Jones L. The alcohol harm paradox: using a national survey to explore how alcohol may disproportionately impact health in deprived individuals. BMC Public Health 2016;16:111.
Birch S, Jerrett M, Wilson K, Law M, et al. Heterogeneities in the production of health: smoking, health status and place. Health Policy 2005;72:301-310.
Dowler E. Policy initiatives to address low-income households’ nutritional needs in the UK. Proceedings of the Nutrition Society 2008;67:289–300. doi:10.1017/S0029665108008586.
Hart CL, Gruer L,Watt GCM. Cause specific mortality, social position, and obesity among women who had never smoked: 28 year cohort study. BMJ 2011;342.d3785. doi: https://doi.org/10.1136/bmj.d3785.
Nandi A, Glymour MM, Subramanian SV. Association Among Socioeconomic Status, Health Behaviors, and All-Cause Mortality in the United States. Epidemiology 2014;25:170-177. doi: 10.1097/EDE.0000000000000038.
Pampel FC, Rogers RG. Socioeconomic status, smoking, and health: a test of competing theories of cumulative advantage. J Health Soc Behav 2004;45:306-321.
Stringhini S, Dugravot A, Shipley M, et al. Health Behaviours, Socioeconomic Status, and Mortality: Further Analyses of the British Whitehall II and the French GAZEL Prospective Cohorts. PLoS Med 2011;8:e1000419.
The authors of this paper aimed to investigate whether and to what extent consumption by beverage type, BMI, smoking and other factors explain inequalities in alcohol-related harm evidenced though wholly attributable alcohol-related hospital admissions (ARHA). They especially evaluated levels of socio-economic status (SES) as determined by geographic and individual levels of deprivation. They based their analyses on data from 11,038 respondents to the Welsh Health Survey who answered questions on their health and lifestyle.
The results of this study support previous research on how factors associated with poor socio-economic status relate to poor health outcomes. It provides added information by providing an expanded definition for deprivation (assessed from geographic area of residence as well as from personal data), and for potential differences by type of beverage consumed. The definition of deprivation included many factors that, together, seem to be the primary reason why poorer, less educated people seem to have more adverse health effects that have been attributed to alcohol consumption.
Specifically, this study shows that the amount of alcohol consumed by people with low SES may be only a minor factor in explaining their increase in adverse effects that are usually classified as being “alcohol-related.” In this study, the frequency of alcohol consumption explained only 4% of the greater number of “wholly attributable alcohol-related hospital admissions.” Co-existing other behaviors and conditions, especially smoking, obesity, living in a deprived environment, poor health status, and mental health, appeared to have much greater effect. Thus, the conclusions of this study suggest that even for what are referred to as “wholly attributable .alcohol-related hospital admissions,” other factors associated with low SES may be more important than alcohol consumption itself as determinants of such adverse health outcomes.
* * * * * * *
Comments on this critique by the International Scientific Forum on Alcohol Research were provided by the following members:
Yuqing Zhang, MD, DSc, Clinical Epidemiology, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA.
David Van Velden, MD, Dept. of Pathology, Stellenbosch University, Stellenbosch, South Africa
Pierre-Louis Teissedre, PhD, Faculty of Oenology–ISVV, University Victor Segalen Bordeaux 2, Bordeaux, France
Dag S. Thelle, MD, PhD, Department of Biostatistics, Institute of Basic Medical Sciences, University of Oslo, Norway; Section for Epidemiology and Social Medicine, Sahlgrenska Academy, University of Gothenburg, Sweden
Arne Svilaas, MD, PhD, general practice and lipidology, Oslo University Hospital, Oslo, Norway
Creina Stockley, PhD, MSc Clinical Pharmacology, MBA; Adjunct Senior Lecturer at the University of Adelaide, Australia
Erik Skovenborg, MD, specialized in family medicine, member of the Scandinavian Medical Alcohol Board, Aarhus, Denmark
Fulvio Mattivi, MSc, CAFE – Center Agriculture Food Environment, University of Trento, via E. Mach 1, San Michele all’Adige, Italy
Imke Janssen, PhD, Department of Preventive Medicine, Rush University Medical Centre, Chicago, IL, USA.
Harvey Finkel, MD, Hematology/Oncology, Retired (Formerly, Clinical Professor of Medicine, Boston University Medical Center, Boston, MA, USA)
R. Curtis Ellison, MD, Professor of Medicine, Section of Preventive Medicine & Epidemiology, Boston University School of Medicine, Boston, MA, USA
Giovanni de Gaetano, MD, PhD, Department of Epidemiology and Prevention, IRCCS Istituto Neurologico Mediterraneo NEUROMED, Pozzilli, Italy (with invited input from colleague, Augusto Di Castelnuovo)