Estimating the number of alcohol-attributable deaths: methodological issues and illustration with French data for 2006

Methodological guidelines for the estimation of attributable mortality using a prevalence-based method: The STREAMS-P tool

BACKGROUND

There is evidence of strong links between exposure to different risk factors and life-threatening diseases. Assessing the burden of a risk factor on the population's mortality due to a given disease provides a clear picture of these links. The estimation of attributable mortality to a risk factor is the most widely used procedure for doing this. Although different methods are available to estimate attributable mortality, the prevalence-based methodology is the most frequent. The main objective of this paper is to develop guidelines and checklists to STrengthen the design and REporting of Attributable-Mortality Studies using a Prevalence-based method (STREAMS-P) and also to assess the quality of an already published study which uses this methodology.

METHODS

The design of the guideline and checklists has been done in two phases. A development phase, where we set recommendations based on the review of the literature; and a validation phase, where we validated our recommendations against other published studies that have estimated attributable mortality using a prevalence-based method.

RESULTS

We have developed and tested a guideline that includes the information required to perform a prevalence-based attributable mortality study to a given risk factor; a checklist of aspects that should be present when a report or a paper on attributable mortality is written or interpreted and a checklist of quality control criteria for reports or papers estimating attributable mortality.

CONCLUSION

To our knowledge, the STREAMS-P is the first set of criteria specifically created to assess the quality of such studies and it could be valuable for authors and readers interested in performing attributable mortality studies or interpreting their reliability.

Future Alcohol-Attributable Mortality in France Using a Novel Generalizable Age-Period-Cohort Projection Methodology

AIM

To forecast age- and sex-specific alcohol-attributable mortality in France for the period 2015-2050 using a novel generalizable methodology that includes different scenarios regarding period and cohort change.

METHODS

For the French national population aged 25-90 years (1979-2014), we estimated alcohol-attributable mortality by mortality from the main causes of death wholly attributable to alcohol, plus liver cirrhosis mortality. We modelled sex-specific alcohol-attributable mortality by adjusting for age, period and birth cohort. We forecasted the model parameters to obtain future age- and sex-specific alcohol-attributable mortality up until 2050 using a conventional baseline, scenario I (favourable period change) and scenario II (unfavourable cohort change).

RESULTS

Alcohol-attributable mortality is clearly declining in France, with the decline decelerating from 1992 onwards. In 2014, the age-standardized alcohol-attributable mortality rates, in deaths per 100,000, were 34.7 among men and 9.9 among women. In 2050, the estimated rates are between 10.5 (prediction interval: 7.6-14.4; scenario I) and 17.6 (13.1-23.7; scenario II) among men, and between 1.1 (0.7-1.7; scenario I) and 1.8 (1.2-2.9; scenario II) among women; which implies declines of 58% for men and 84% for women (baseline).

CONCLUSION

Alcohol-attributable mortality in France is expected to further decline in the coming decades, accompanied by age pattern changes. However, France's levels are not expected to reach the current lower levels in Italy and Spain for 15 years or more. Our results point to the value of implementing preventive policy measures that discourage alcohol consumption among people of all ages, but especially among adolescents.

Proportion of cancer cases and deaths attributable to alcohol consumption by US state, 2013-2016

BACKGROUND

Alcohol consumption is an established risk factor for several cancer types, but there are no contemporary published estimates of the state-level burden of cancer attributed to alcoholic beverage consumption. Such estimates are needed to inform public policy and cancer control efforts. We estimated the proportion and number of incident cancer cases and cancer deaths attributable to alcohol consumption by sex in adults aged ≥30 years in all 50 states and the District of Columbia in 2013-2016.

METHODS

Age-, sex-, and state-specific cancer incidence and mortality data (2013-2016) were obtained from the US Cancer Statistics database. State-level, self-reported age and sex stratified alcohol consumption prevalence was estimated using the 2003-2006 Behavioral Risk Factor Surveillance System surveys and adjusted with state sales data.

RESULTS

The proportion of alcohol-attributable incident cancer cases ranged from 2.9 % (95 % confidence interval: 2.7 %-3.1 %) in Utah to 6.7 % (6.4 %-7.0 %) in Delaware among men and women combined, from 2.7 % (2.5 %-3.0 %) in Utah to 6.3 % (5.9 %-6.7 %) in Hawaii among men, and from 2.7 % (2.4 %-3.0 %) in Utah to 7.7 % (7.2 %-8.3 %) in Delaware among women. The proportion of alcohol-attributable cancer deaths also varied considerably across states: from 1.9 % to 4.5 % among men and women combined, from 2.1% to 5.0% among men, and from 1.4 % to 4.4 % among women. Nationally, alcohol consumption accounted for 75,199 cancer cases and 18,947 cancer deaths annually.

CONCLUSION

Alcohol consumption accounts for a considerable proportion of cancer incidence and mortality in all states. Implementing state-level policies and cancer control efforts to reduce alcohol consumption could reduce this cancer burden.

The American Cancer Society 2035 challenge goal on cancer mortality reduction

A summary evaluation of the 2015 American Cancer Society (ACS) challenge goal showed that overall US mortality from all cancers combined declined 26% over the period from 1990 to 2015. Recent research suggests that US cancer mortality can still be lowered considerably by applying known interventions broadly and equitably. The ACS Board of Directors, therefore, commissioned ACS researchers to determine challenge goals for reductions in cancer mortality by 2035. A statistical model was used to estimate the average annual percent decline in overall cancer death rates among the US general population and among college-educated Americans during the most recent period. Then, the average annual percent decline in the overall cancer death rates of college graduates was applied to the death rates in the general population to project future rates in the United States beginning in 2020. If overall cancer death rates from 2020 through 2035 nationally decline at the pace of those of college graduates, then death rates in 2035 in the United States will drop by 38.3% from the 2015 level and by 54.4% from the 1990 level. On the basis of these results, the ACS 2035 challenge goal was set as a 40% reduction from the 2015 level. Achieving this goal could lead to approximately 1.3 million fewer cancer deaths than would have occurred from 2020 through 2035 and 122,500 fewer cancer deaths in 2035 alone. The results also show that reducing the prevalence of risk factors and achieving optimal adherence to evidence-based screening guidelines by 2025 could lead to a 33.5% reduction in the overall cancer death rate by 2035, attaining 85% of the challenge goal.

Comparison of different approaches for estimating age-specific alcohol-attributable mortality: The cases of France and Finland

Background

Accurate estimates of the impact of alcohol on overall and age-specific mortality are crucial for formulating health policies. However, different approaches to estimating alcohol-attributable mortality provide different results, and a detailed comparison of age-specific estimates is missing.

Methods

Using data on cause of death, alcohol consumption, and relative risks of mortality at different consumption levels, we compare eight estimates of sex- and age-specific alcohol-attributable mortality in France (2010) and Finland (2013): five estimates using cause-of-death approaches (with one accounting for contributory causes), and three estimates using attributable fraction (AF) approaches.

Results

AF-related approaches and the approach based on alcohol-related underlying and contributory causes of death provided estimates of alcohol-attributable mortality that were twice as high as the estimates found using underlying cause-of-death approaches in both countries and sexes. The differences across the methods were greatest among older age groups An inverse U-shape in age-specific alcohol-attributable mortality (peaking at around age 65) was observed for cause-of-death approaches, with this shape being more pronounced in Finland. AF-related approaches resulted in different estimates at older ages: i.e., mortality was found to increase with age in France; whereas in Finland mortality estimates depended on the underlying assumptions regarding the effects of alcohol consumption on cardiovascular mortality.

Conclusions

While the most detailed approaches (i.e., the AF-related approach and the approach that includes underlying and contributory causes) are theoretically able to provide more accurate estimates of alcohol-attributable mortality, they–especially the AF approaches- depend heavily on data availability and quality. To enhance the reliability of alcohol-attributable mortality estimates, data quality for older age groups needs to be improved.

Proportion and number of cancer cases and deaths attributable to potentially modifiable risk factors in the United States

Contemporary information on the fraction of cancers that potentially could be prevented is useful for priority setting in cancer prevention and control. Herein, the authors estimate the proportion and number of invasive cancer cases and deaths, overall (excluding nonmelanoma skin cancers) and for 26 cancer types, in adults aged 30 years and older in the United States in 2014, that were attributable to major, potentially modifiable exposures (cigarette smoking; secondhand smoke; excess body weight; alcohol intake; consumption of red and processed meat; low consumption of fruits/vegetables, dietary fiber, and dietary calcium; physical inactivity; ultraviolet radiation; and 6 cancer-associated infections). The numbers of cancer cases were obtained from the Centers for Disease Control and Prevention (CDC) and the National Cancer Institute; the numbers of deaths were obtained from the CDC; risk factor prevalence estimates were obtained from nationally representative surveys; and associated relative risks of cancer were obtained from published, large-scale pooled analyses or meta-analyses. In the United States in 2014, an estimated 42.0% of all incident cancers (659,640 of 1570,975 cancers, excluding nonmelanoma skin cancers) and 45.1% of cancer deaths (265,150 of 587,521 deaths) were attributable to evaluated risk factors. Cigarette smoking accounted for the highest proportion of cancer cases (19.0%; 298,970 cases) and deaths (28.8%; 169,180 deaths), followed by excess body weight (7.8% and 6.5%, respectively) and alcohol intake (5.6% and 4.0%, respectively). Lung cancer had the highest number of cancers (184,970 cases) and deaths (132,960 deaths) attributable to evaluated risk factors, followed by colorectal cancer (76,910 cases and 28,290 deaths). These results, however, may underestimate the overall proportion of cancers attributable to modifiable factors, because the impact of all established risk factors could not be quantified, and many likely modifiable risk factors are not yet firmly established as causal. Nevertheless, these findings underscore the vast potential for reducing cancer morbidity and mortality through broad and equitable implementation of known preventive measures. CA Cancer J Clin 2018;68:31-54. © 2017 American Cancer Society.

A new methodological approach to adjust alcohol exposure distributions to improve the estimation of alcohol-attributable fractions

BACKGROUND AND AIMS

To assess the burden of excessive alcohol use, researchers estimate alcohol-attributable fractions (AAFs) routinely. However, under-reporting in survey data can bias these estimates. We present an approach that adjusts for under-reporting in the estimation of AAFs, particularly within subgroups. This framework is a refinement of a previous method conducted by Rehm et al.

METHODS

We use a measurement error model to derive the 'true' alcohol distribution from a 'reported' alcohol distribution. The 'true' distribution leverages per-capita sales data to identify the distribution average and then identifies the shape of the distribution with self-reported survey data. Data are from the National Alcohol Survey (NAS), the National Household Survey on Drug Abuse (NHSDA) and the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC). We compared our approach with previous approaches by estimating the AAF of female breast cancer cases.

RESULTS

Compared with Rehm et al.'s approach, our refinement performs similarly under a gamma assumption. For example, among females aged 18-25 years, the two approaches produce estimates from NHSDA that are within a percentage point. However, relaxing the gamma assumption generally produces more conservative evidence. For example, among females aged 18-25 years, estimates from NHSDA based on the best-fitting distribution are only 19.33% of breast cancer cases, which is a much smaller proportion than the gamma-based estimates of approximately 28%.

CONCLUSIONS

A refinement of Rehm et al.'s approach to adjusting for underreporting in the estimation of alcohol-attributable fractions provides more flexibility. This flexibility can avoid biases associated with failing to account for the underlying differences in alcohol consumption patterns across different study populations. Comparisons of our refinement with Rehm et al.'s approach show that results are similar when a gamma distribution is assumed. However, results are appreciably lower when the best-fitting distribution is chosen versus gamma-based results.

The role of birth cohorts in long-term trends in liver cirrhosis mortality across eight European countries

BACKGROUND AND AIMS

Understanding why inequalities in alcohol-related mortality trends by sex and country exist is essential for developing health policies. Birth cohort effects, indicative of differences by generation in drinking, have rarely been studied. This study estimated the relative contributions of birth cohorts to liver cirrhosis mortality trends and compared sex- and country-specific cohort patterns across eight European countries.

DESIGN

Time-series analysis of population-level mortality data.

SETTING

Austria, Finland, Hungary, Italy, the Netherlands, Poland, Spain and Sweden; 1950-2011.

PARTICIPANTS

National populations aged 15-94 years.

MEASUREMENTS

We modelled country- and sex-specific liver cirrhosis mortality (from national vital registers) adjusting for age, period and birth cohort.

FINDINGS

Birth cohorts (adjusted for age and period) made statistically significant contributions to liver cirrhosis mortality in all countries and for both sexes (P < 0.001), and more so among women (average contribution to deviance reduction of 38.8%) than among men (17.4%). The observed cohort patterns were statistically different between all but two country pairs (P < 0.001). Sex differences existed overall (P < 0.001), but not in the majority of countries (P > 0.999). Visual inspection of birth cohort patterns reveals birth cohorts at higher risk of liver cirrhosis mortality.

CONCLUSIONS

The inclusion of the birth cohort dimension improves the understanding of alcohol-attributable mortality trends in Europe. Birth cohorts at higher risk of liver cirrhosis mortality were born during 1935-49 in Sweden and Finland, around 1950 in Austria and the Netherlands and 1960 or later in Hungary, Italy, Poland and Spain.

A narrative review of alcohol consumption as a risk factor for global burden of disease

Since the original Comparative Risk Assessment (CRA) for alcohol consumption as part of the Global Burden of Disease Study for 1990, there had been regular updates of CRAs for alcohol from the World Health Organization and/or the Institute for Health Metrics and Evaluation. These studies have become more and more refined with respect to establishing causality between dimensions of alcohol consumption and different disease and mortality (cause of death) outcomes, refining risk relations, and improving the methodology for estimating exposure and alcohol-attributable burden. The present review will give an overview on the main results of the CRAs with respect to alcohol consumption as a risk factor, sketch out new trends and developments, and draw implications for future research and policy.

Mortality among homeless people in France, 2008-10

BACKGROUND

Studies in various countries have shown that homeless people have high mortality levels. The aims of this study concerning the French population were to investigate mortality among the homeless and to study their causes of death in comparison to those of the general population.

METHODS

A representative sample of 1145 homeless deaths registered by an association was matched to the national database of medical causes of death using common descriptive variables. Log-binomial regression was used to compare mortality among the homeless to that of the general population. Multiple imputation was used to manage missing causes of deaths.

RESULTS

Out of the 1145 registered homeless deaths, 693 were matched to the causes of death database. Homeless deaths were young (average age: 49). Overall, homeless deaths were slightly more frequent during winter. Among all deaths, the probability of being homeless was higher when dying from hypothermia (RR = 6.4), alcohol-related deaths (RR = 1.7), mental disorders, diseases of the digestive and circulatory systems, and undetermined causes (RR from 1.5 to 3.7).

CONCLUSION

The homeless died at 49 years old on average compared with 77 in the general population in 2008-10. The health of homeless people should be considered not only in winter periods or in terms of alcohol- or cold-related conditions. This study also highlights the need for more precise data to estimate the mortality risks of the homeless in France.

Estimates of alcohol-related oesophageal cancer burden in Japan: systematic review and meta-analyses

OBJECTIVE

To refine estimates of the burden of alcohol-related oesophageal cancer in Japan.

METHODS

We searched PubMed for published reviews and original studies on alcohol intake, aldehyde dehydrogenase polymorphisms, and risk for oesophageal cancer in Japan, published before 2014. We conducted random-effects meta-analyses, including subgroup analyses by aldehyde dehydrogenase variants. We estimated deaths and loss of disability-adjusted life years (DALYs) from oesophageal cancer using exposure distributions for alcohol based on age, sex and relative risks per unit of exposure.

FINDINGS

We identified 14 relevant studies. Three cohort studies and four case-control studies had dose-response data. Evidence from cohort studies showed that people who consumed the equivalent of 100 g/day of pure alcohol had an 11.71 fold, (95% confidence interval, CI: 2.67-51.32) risk of oesophageal cancer compared to those who never consumed alcohol. Evidence from case-control studies showed that the increase in risk was 33.11 fold (95% CI: 8.15-134.43) in the population at large. The difference by study design is explained by the 159 fold (95% CI: 27.2-938.2) risk among those with an inactive aldehyde dehydrogenase enzyme variant. Applying these dose-response estimates to the national profile of alcohol intake yielded 5279 oesophageal cancer deaths and 102,988 DALYs lost - almost double the estimates produced by the most recent global burden of disease exercise.

CONCLUSION

Use of global dose-response data results in an underestimate of the burden of disease from oesophageal cancer in Japan. Where possible, national burden of disease studies should use results from the population concerned.

Small area associations between social context and alcohol-attributable mortality in a middle income country

BACKGROUND

Little is known about the association between alcohol-attributable mortality and small area socioeconomic variables when considering causes both wholly and partially attributable to alcohol.

METHODS

An ecological study was conducted of the entire Chilean population aged 15 and older in 345 municipalities nationwide between 2004 and 2009. Deaths were attributed to alcohol consumption either wholly or partially, along with the estimated attributable fractions for each specified cause. Each municipality was characterized according to its average income and educational attainment. Estimates of the ecological associations were produced using a hierarchical Bayesian model, separating out deaths caused by alcohol and dividing them into seven groups of causes.

RESULTS

Alcohol-attributable mortality risk showed an inverse association with income and education at the ecological level. A one-quintile increase in income was associated with an average decrease in risk of 10% (CI 95%: 10-20%) for cardiovascular deaths, 8% (6-10%) for intentional injuries and 7% (3-11%) for unintentional injuries. No associations were found between deaths due to cancers and other causes with income and education.

CONCLUSIONS

Municipalities with lower income and education have higher risk of alcohol-attributable mortality in Chile.

Estimation of usual occasion-based individual drinking patterns using diary survey data

BACKGROUND

In order to successfully address excessive alcohol consumption it is essential to have a means of measuring the drinking patterns of a nation. Owing to the multi-dimensional nature of drinking patterns, usual survey methods have their limitations. The aim of this study was to make use of extremely detailed diary survey data to demonstrate a method of combining different survey measures of drinking in order to reduce these limitations.

METHODS

Data for 1724 respondents of the 2000/01 National Diet and Nutrition Survey was used to obtain a drinking occasion dataset, by plotting the respondent's blood alcohol content over time. Drinking frequency, level and variation measures were chosen to characterise drinking behaviour and usual behaviour was estimated via statistical methods.

RESULTS

Complex patterns in drinking behaviour were observed amongst population subgroups using the chosen consumption measures. The predicted drinking distribution combines diary data equivalent coverage with a more accurate proportion of non-drinkers.

CONCLUSIONS

This statistical analysis provides a means of obtaining average consumption measures from diary data and thus reducing the main limitation of this type of data for many applications. We hope that this will facilitate the use of such data in a wide range of applications such as risk modelling, especially for acute harms, and burden of disease studies.

Alcohol-attributable cancer deaths and years of potential life lost in the United States

OBJECTIVES

Our goal was to provide current estimates of alcohol-attributable cancer mortality and years of potential life lost (YPLL) in the United States.

METHODS

We used 2 methods to calculate population-attributable fractions. We based relative risks on meta-analyses published since 2000, and adult alcohol consumption on data from the 2009 Alcohol Epidemiologic Data System, 2009 Behavioral Risk Factor Surveillance System, and 2009-2010 National Alcohol Survey.

RESULTS

Alcohol consumption resulted in an estimated 18,200 to 21,300 cancer deaths, or 3.2% to 3.7% of all US cancer deaths. The majority of alcohol-attributable female cancer deaths were from breast cancer (56% to 66%), whereas upper airway and esophageal cancer deaths were more common among men (53% to 71%). Alcohol-attributable cancers resulted in 17.0 to 19.1 YPLL for each death. Daily consumption of up to 20 grams of alcohol (≤ 1.5 drinks) accounted for 26% to 35% of alcohol-attributable cancer deaths.

CONCLUSIONS

Alcohol remains a major contributor to cancer mortality and YPLL. Higher consumption increases risk but there is no safe threshold for alcohol and cancer risk. Reducing alcohol consumption is an important and underemphasized cancer prevention strategy.

Modeling the impact of alcohol dependence on mortality burden and the effect of available treatment interventions in the European Union

Alcohol consumption is a major risk factor for the burden of disease, and Alcohol Dependence (AD) is the most important disorder attributable to this behavior. The objective of this study was to quantify mortality associated with AD and the potential impact of treatment. For the EU countries, for the age group 15-64 years, mortality attributable to alcohol consumption in general, to heavy drinking, and to AD were estimated based on the latest data on exposure and mortality. Potential effects of AD treatment were modeled based on Cochrane and other systematic reviews of the effectiveness of the best known and most effective interventions. In the EU 88.9% of men and 82.1% of women aged 15-64 years were current drinkers; and 15.3% of men and 3.4% of women in this age group were heavy drinkers. AD affected 5.4% of men and 1.5% of women. The net burden caused by alcohol consumption was 1 in 7 deaths in men and 1 in 13 deaths in women. The majority of this burden was due to heavy drinking (77%), and 71% of this burden was due to AD. Increasing treatment coverage for the most effective treatments to 40% of all people with AD was estimated to reduce alcohol-attributable mortality by 13% for men and 9% for women (annually 10,000 male and 1700 female deaths avoided). Increasing treatment rates for AD was identified as an important issue for future public health strategies to reduce alcohol-attributable harm and to complement the current focus of alcohol policy.

Combining best evidence: a novel method to calculate the alcohol-attributable fraction and its variance for injury mortality

BACKGROUND

The alcohol-attributable fraction for injury mortality is defined as the proportion of fatal injury that would disappear if consumption went to zero. Estimating this fraction has previously been based on a simplistic view of drinking and associated risk. This paper develops a new way to calculate the alcohol-attributable fraction for injury based on different dimensions of drinking, mortality data, experimental data, survey research, new risk scenarios, and by incorporating different distributions of consumption within populations. For this analysis, the Canadian population in 2005 was used as the reference population.

METHODS

Binge drinking and average daily consumption were modeled separately with respect to the calculation of the AAF. The acute consumption risk was calculated with a probability-based method that accounted for both the number of binge drinking occasions and the amount of alcohol consumed per occasion. The average daily consumption was computed based on the prevalence of daily drinking at various levels. These were both combined to get an overall estimate. 3 sensitivity analyses were performed using different alcohol consumption parameters to test the robustness of the model. Calculation of the variance to generate confidence limits around the point estimates was accomplished via Monte Carlo resampling methods on randomly generated AAFs that were based on the distribution and prevalence of drinking in the Canadian population.

RESULTS

Overall, the AAFs decrease with age and are significantly lower for women than men across all ages. As binge drinking increases, the injury mortality AAF also increases. Motor vehicle collisions show the largest relative increases in AAF as alcohol consumption is increased, with over a 100% increase in AAF from the lowest to highest consumption category. Among non-motor vehicle collisions, the largest change in total AAF occurred both for homicide and other intentional injuries at about a 15% increase in the AAF from the lowest to the highest binge consumption scenarios.

CONCLUSIONS

This method combines the best available evidence to generate new alcohol-attributable fractions for alcohol-attributable injury mortality. Future research is needed to refine the risk function for non-motor vehicle injury types and to investigate potential interactions between binge drinking and average volume of alcohol consumption.