Article Text
Abstract
Background During the past decades, social inequality in mortality has increased in several countries, including Denmark. Modifiable risk factors, such as smoking and harmful alcohol consumption, have been suggested to moderate the association between socioeconomic position and health-related outcomes. The present study aims to investigate the contribution of smoking- and alcohol-related deaths to the trends in educational inequality in mortality in Denmark 1995–2019 among individuals aged 30–74 years.
Methods Nationwide data on mortality and highest attained educational level divided into quartiles were derived from administrative registers. Alcohol-related mortality was directly estimated using information on alcohol-related deaths from death certificates. Smoking-related mortality was indirectly estimated using the Peto-Lopez method. The contribution of smoking- and alcohol-related deaths to the social inequality gap in mortality 1995–2019 was calculated.
Results Alongside a decrease in all-cause mortality in Denmark 1995–2019, absolute differences in the mortality rate (per 100 000 person-year) between the lowest and the highest educational quartile increased from 494 to 607 among men and from 268 to 376 among women. Among both men and women, smoking- and alcohol-related deaths explained around 60% of the social inequality in mortality and around 50% of the increase in mortality inequality.
Conclusion Smoking and harmful alcohol consumption continue to be important risk factors and causes of social inequality in mortality, with around half of the increase in Denmark 1995–2019 being attributable to smoking- and alcohol-related deaths. Future healthcare planning and policy development should aim at reducing social inequality in modifiable health risk behaviours and their negative consequences.
- MORTALITY
- EPIDEMIOLOGY
- Health inequalities
- PUBLIC HEALTH
Data availability statement
No data are available. Not applicable.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
During the past decades, all-cause as well as smoking-related and alcohol-related mortality have decreased in many countries, including Denmark. However, this improvement is not evenly distributed within society, leaving individuals with lower socioeconomic position less advantaged than those with a higher position. This has resulted in social inequality in both all-cause mortality, and smoking- and alcohol-related mortality, and previous studies have indicated an increasing social inequality in all-cause mortality attributable to smoking- and alcohol-related deaths.
WHAT THIS STUDY ADDS
This study adds updated results on the trends in social inequality in mortality in Denmark 1995–2019. Specifically, it estimates the contribution of smoking- and alcohol-related deaths on the trends in social inequality in mortality in the study period. The results show that smoking- and alcohol-related deaths can explain around 60% of the social inequality in mortality and around 50% of the increased mortality gap during the past decades.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
Even though a general decrease in both all-cause, smoking- and alcohol-related mortality has been observed during the past decades, social inequality in mortality continues to increase. Accordingly, as smoking and harmful alcohol consumption are modifiable risk factors, social inequality in health may be reduced by targeting social differences in such behaviours. Moreover, preventive health screening programmes or socially differentiated counselling in relation to hospital treatment might be considered to reduce social inequality in health at a societal level.
Background
The strong influence of socioeconomic factors on health, morbidity, and mortality is well documented.1–4 In addition, even though a general increase in life expectancy has been observed in many developed countries, including Denmark, this improvement is not evenly distributed within society. This has resulted in an increasing social inequality in mortality over the past decades.4–14 Consequently, reducing social inequalities in health has received increasing political attention and is now among the most important and prioritised targets in public health policies.15
Social inequality in health is partly explained by an unequal exposure to various environmental risk factors and means that less advantaged groups have a significantly higher risk of morbidity and mortality.16 The causes of social inequality are complex, multifactorial and related to factors that are both established early in life and persist throughout the life course.17 18 However, studies have suggested that modifiable risk factors, such as smoking and harmful alcohol consumption, may moderate the association between social position and health-related outcomes, for example, mortality.19–21 Although both risk factors have consistently been found to be causally associated with an increased risk of morbidity and mortality, latency periods are rather long, for example, around 10–30 years for lung cancer incidence.22 Accordingly, preventive strategies aiming at reducing social inequality in mortality by targeting smoking and harmful alcohol consumption need to keep this in mind in their evaluations.
Previous studies have examined both the separate23–30 and joint31 32 contribution of smoking- and alcohol-related deaths to the trends in social inequality in all-cause mortality or life expectancy. Studies examining the separate contribution show relatively inconsistent patterns, with an increasing contribution of smoking-related deaths to mortality inequality, especially among women, along with a stagnation or a decrease among men.23 25–28 Conflicting results are observed regarding the contribution of alcohol-related deaths among both men and women.24 25 29 30 The joint contribution of smoking- and alcohol-related deaths has been found to widen social inequality in mortality from the 1980s until the late 2000s.31 32 However, to our knowledge, no study has investigated the joint contribution after this period and until today.
The aim of this study was to investigate the contribution of smoking- and alcohol-related deaths to the trends in educational inequality in all-cause mortality in Denmark 1995–2019 among individuals aged 30–74 years.
Methods
This study is a register-based study using individual-level data from nationwide administrative registers in Denmark. Thus, information on sex and age was derived from the Danish Civil Registration System,33 whereas information on the highest attained educational level and causes of death was derived from the Danish Education Registers34 and the Danish Register of Causes of Death,35 respectively.
In Denmark, every individual with a permanent residence has a unique personal identification number (CPR number),33 and by using CPR numbers, linkage between information from different registers is possible.
Study population
In this study, the study population was constituted by every individual with a Danish CPR number aged 30–74 year between 1 January 1995 and 31 December 2019. Individuals younger than 30 years or 75 years or older were excluded because of uncertainties regarding educational attainments.
Highest attained educational level
Highest attained educational level was used as an indicator for social position and obtained by linkage to the Danish Education Registers.34 The register is updated with new information in October every year; thus, in this study, educational information covers the period from 1 October 1994 to 1 October 2018. As educational levels have generally increased in the Danish population during the study period, educational quartiles were constructed for each year in the study period to ensure equal sizes of each educational group over time. These educational quartiles were based on continuous educational data and defined separately by sex and 5-year age groups to mirror the applied approach of the mortality calculations. Social inequality was expressed as the mortality differences between the lowest and the highest educational quartiles, with the remaining two quartiles (‘low’ and ‘high’) placed in between. In online supplemental material A, educational distributions according to the International Standard Classification of Education (ISCED) from 2011 are presented across the study period.
Supplemental material
Mortality
All deaths that occur in Denmark are registered in the Danish Register of Causes of Death, which dates back to 1970 in its current form.35 All-cause mortality covers all deaths with no stratification on disease groups. For all registered deaths, the death certificate includes the underlying cause of death and any contributing causes of deaths, which are classified according to the International Classification of Diseases, 10th Edition (ICD-10) from 1994.35 In this study, data on causes of death between 1 January 1995 and 31 December 2019 were used.
In analyses of all-cause mortality and smoking-related mortality, only the underlying cause of death was included. In contrast, both underlying and contributing causes of death were included for alcohol-related mortality. Alcohol-related deaths were defined as an underlying or contributing cause of death with at least one of the following ICD-10 codes: Alcoholism and alcohol psychosis (F10), cirrhosis of the liver (K70, K74), pancreatitis (K85–K86), and alcohol poisoning (X45, X65, Y15). Smoking-related deaths are more complex to define, as death certificates do not contain information about whether the death was or could have been caused by smoking. A frequently used indirect method to estimate smoking-related mortality was developed by Peto et al.36 The calculations in this method, referred to as the Peto-Lopez method, are based on aetiological fractions, which uses estimates of lung cancer mortality among never-smokers to estimate the contribution of smoking on the mortality of other site-specific cancer diagnoses, chronic obstructive pulmonary disease, other respiratory diseases, cardiovascular diseases and other natural causes of death. To ensure that the smoking-related mortality is not overestimated, the Peto-Lopez method adjust for unmeasured confounding, and in this study, a 30% reduction of the excess risk was used. It is assumed that no smoking-related deaths occur before age 35, and that no deaths caused by cirrhosis of the liver or occurring by accident, suicide or homicide are related to smoking.
Results are presented as mortality rates per 100 000 person-years, calculated in 5-year averaged groups (1995–1999, 2015–2019), and stratified by sex, highest attained educational quartiles and 5-year age groups. Calculations were conducted by dividing the total number of deaths in the period with the total time at risk. As a measure of social inequality, the absolute difference in mortality rates in the lowest and in the highest educational quartile group, respectively, is presented. For internationally comparative purposes, calculations of social inequality in all-cause mortality based on educational level according ISCED are presented as age-standardised rates per 100 000 person-years in online supplemental material B, along with risk ratio (RR) and rate differences between ISCED levels 0–2 and 5–8. Trends in social inequality in all-cause mortality calculated as age-standardised differences in the Slope Index of Inequality (SII) using ISCED levels are presented in online supplemental material C. These supplementary analyses support the findings in our main analyses presented in the result section.
Supplemental material
Supplemental material
The contribution of smoking- and alcohol-related deaths to the trends in social inequality in all-cause mortality is presented separately as the absolute difference in percentages between the first period (1995–1999) and the last period (2015–2019). The European standard population was used as the reference population when conducting age-standardised analyses.
Results
As shown in table 1, the total number of deaths decreased between 1995 and 2019 among both men and women in Denmark aged 30–74 years. When stratified by cause of death, a decrease was observed for smoking-related deaths and deaths from other causes. The number of alcohol-related deaths remained relatively stable; however, with some fluctuations during the period and with a decreasing trend among women since 2010. The contribution of smoking- and alcohol-related deaths to the total number of deaths among men decreased from 40.6% in 1995–1999 to 33.9% in 2015–2019, with the largest decrease between 2010–2014 and 2015–2019 (−4.3 percentage points). Among women, the contribution remained relatively stable (35%–38%). Consequently, the contribution of deaths of other causes increased among men between 1995 and 2019 but remained relatively stable among women.
In figure 1A–C, the age-standardised all-cause, smoking-related and alcohol-related mortality rates, respectively, are presented, according to educational quartiles and period. Across all educational groups and among both men and women, a decrease in all-cause, smoking-related and alcohol-related mortality, respectively, is observed during the period, however, with a few exceptions. For example, the alcohol-related mortality increased between 2000 and 2009 among women in the lowest educational quartile. Throughout the period and across causes of death, the highest mortality rates are observed among individuals in the lowest educational quartile and lowest in the highest educational quartile. The two intermediate quartiles are placed in between the lowest and highest quartiles.
The educational gap, that is, the difference between the lowest and the highest educational quartile, in all-cause, smoking-related and alcohol-related mortality rates, respectively, has generally widened during the period among both men and women. This is due to a more pronounced decrease among individuals in the highest educational quartile. Specifically, between 1995–1999 and 2015–2019 the absolute difference in all-cause mortality rates increased from 494 to 607 deaths per 100 000 person-years among men, and from 268 to 376 deaths per 100 000 person-years among women. For smoking-related mortality, the absolute difference in mortality rates between the lowest and the highest educational quartile increased from 226 to 255 deaths per 100 000 person-years among men between 1995–1999 and 2015–2019, and from 165 to 197 deaths per 100 000 person-years among women. However, during the period 1995–2019, some fluctuations are observed, for example, among men where a large inequality increase is seen from 1995 to 2014, followed by a decrease. Among women, an increased inequality is seen from 1995 to 2009, after which the difference between the lowest and the highest educational quartile remained relatively stable. For alcohol-related mortality, the absolute difference in mortality rates between the lowest and the highest educational quartile increased from 90 to 117 deaths per 100 000 person-years among men between 1995–1999 and 2015–2019, and from 12 to 37 deaths per 100 000 person-years among women. Among both men and women, the increased inequality occurred primarily between 1995 and 2009.
Figure 2 shows the absolute difference in mortality rates among men and women aged 30–74 years between the lowest and the highest educational quartile from 1995 to 2019, according to cause of death group and period. The inequality has increased for most causes of death during the period, with other causes exhibiting the largest inequality increase.
The proportional contributions of these causes of death on social inequality and trends in social inequality in all-cause mortality between 1995 and 2019 are presented in table 2. Among both men and women, approximately 60% of the social inequality in mortality can be attributed to smoking- and alcohol-related deaths, irrespective of period. Online supplemental material D presents corresponding analyses based on SII and shows similar patterns. Between 1995 and 2019, smoking- and alcohol-related deaths could explain around half of the increase in social inequality mortality in Denmark, with a slightly larger proportion among men (53%) than among women (47%).
Supplemental material
As shown in table 3, an increase in social inequality in mortality between 1995 and 2019 comparable to that of smoking- and alcohol-related deaths is attributable to deaths from other causes (47% among men and 53% among women). Separately, smoking-related deaths contribute to 26% and 29% of the increase in social inequality in mortality among men and women, respectively, whereas 23% and 24% of the increase is attributable to alcohol-related deaths.
Discussion
The present nationwide Danish study among individuals aged 30–74 years examined trends in social inequality in mortality between 1995 and 2019 and the contribution of smoking- and alcohol-related deaths on these trends using the highest attained educational level as an indicator for socioeconomic position. We found that both all-cause mortality, smoking-related mortality and alcohol-related mortality decreased during the period across educational quartiles. However, the decrease was more pronounced among individuals with the highest education compared with those with the lowest education, resulting in a widening inequality gap during the study period. Thus, around 60% of the social inequality in mortality could be attributed to smoking- and alcohol-related deaths between 1995 and 2019, and 50% of the increase in mortality inequality during this period could be attributed to smoking- and alcohol-related deaths among both men and women.
Previous studies have examined the contribution of smoking- and alcohol-related deaths to the trends in social inequality in mortality, but exhibit variations in methodologies, for example, the use of direct or indirect measures of cause-specific mortality, the use of absolute or relative measures for social inequality, age of the study population, study period, and mortality and social position measures.2 23–32 Overall, these methodological differences hamper comparisons across studies. However, keeping these methodological variations in mind, previous studies have generally found an increasing separate contribution of smoking-related deaths on social inequality in mortality or life expectancy, especially among women, along with a stagnation or a decrease in the contribution among men.23 25–28 In contrast, conflicting results emerge on the separate contribution of alcohol-related mortality on these trends.24 25 29 30
We have only been able to identify two studies addressing the joint contribution of smoking- and alcohol-related deaths on trends in social inequality in mortality or life expectancy.31 32 A Finnish study from 2014 by Martikainen et al found 69% and 85% of the increase in social inequality in life expectancy between 1988 and 2007 among men and women, respectively, to be attributable to smoking- and alcohol-related deaths.31 In a Danish study by Koch et al from 2015, 75% and 97% of the increase in social inequality in mortality between 1985 and 2009 among men and women, respectively, could be attributed to smoking- and alcohol-related deaths.32 Accordingly, compared with these studies, the results from our study show that smoking- and alcohol-related deaths contributes proportionally less than previously to the increase in social inequality in mortality.
When comparing results from our study with the corresponding results from the study by Koch et al, which was also based on Danish register-based data, the contribution of smoking-related deaths to social inequality in mortality has markedly decreased among women (from 73% in 1985–2009 to 29% in 1995–2019), but slightly increased among men (from 22% to 26%).32 Further, the contribution of alcohol-related deaths decreased among men (from 35% to 23%), but increased among women (from 16% to 24%). A large increase attributable to other causes of death was observed among both men (from 25% to 47%) and women (from 3% to 53%). Some differences between these two studies should, however, be noted, that is, included age group (≥3032 vs 30–74 years) and origin of educational data among individuals aged ≥65 years (imputation32 vs full information).
Comparisons of all-cause mortality across 22 European countries have revealed an average level of social inequality in all-cause mortality in Denmark, both in terms of absolute and relative measures.1 However, among the Nordic countries, Denmark exhibits the highest absolute inequality in all-cause mortality, but the lowest relative inequality.37
This study benefits from major strengths. First, it is based on individual-level data from nationwide registers and thus covers the whole Danish population aged 30–74 years during a period of 25 years, making results generalisable to this age group. Also, using registry data on smoking- and alcohol-related mortality has an advantage over studies based on survey data on health risk behaviours, as they are likely to underestimate true exposure because of for example, recall bias, preferential reporting and non-response. The analyses in our study included full information on the highest attained educational level among all individuals, which is an advantage compared with studies that use imputations among the oldest individuals because of missing educational information. Moreover, another strength is the use of information on the highest attained educational level in completed months as a continuous variable grouped into quartiles. By doing so, it was possible to examine trends across educational groups over time and compute mortality rates in different educational groups, considering a generally increasing educational level in Denmark during the study period. In a study aiming at examining trends in mortality, a limitation of our study is the inclusion of individuals aged 30–74 years, as mortality increases with higher age. However, this age restriction was chosen to achieve full educational information for all included individuals. Even though most deaths occur in the oldest age groups, smoking- and alcohol-related deaths account for a larger proportion of deaths before age 65 years.38 39 Also, educational level has been suggested to lose its discriminatory power at higher ages.30 Accordingly, we assume that the contribution of smoking- and alcohol-related mortality on trends in social inequality is strongest among individuals in an age range included in this study.
Potential limitations include the use of highest attained educational level as an indicator of social position, as other indicators could also have been used, for example, income or labour market participation. However, educational level is generally believed to be the most stable indicator.40 Moreover, we chose to use educational quartile groups to be able to examine trends in social inequality in mortality taking into account a generally increasing educational level in Denmark over time. Other countries may not be able to do so, which may decrease cross-country comparisons. However, supplementary analyses using ISCED may be used for such comparative purposes. The Peto-Lopez method was used to indirectly estimate smoking-related mortality. Although other methods have been developed, for example, Preston-Glei-Wilmoth, Peto-Lopez is still the most commonly applied method.
In conclusion, results from this study show that despite a generally decreasing all-cause mortality in Denmark between 1995 and 2019 across educational quartiles, an increasing social inequality in mortality was concurrently observed. Approximately 50% of this increase can be attributed to smoking- and alcohol-related deaths. Even though structural preventive strategies, for example, tobacco and alcohol control interventions, have been implemented in several countries, including Denmark, they have not yet been able to eliminate the contribution of smoking and harmful alcohol consumption as substantial contributors to an increasing social inequality in mortality. However, we cannot preclude that this may be due to a time lag. Thus, during the entire study period, smoking- and alcohol-related deaths contributed to 60% of the social inequality in all-cause mortality. Accordingly, future healthcare planning and policy development are still encouraged to focus on reducing social inequality in both health risk behaviours and their negative consequences, for example, by targeting smoking and harmful alcohol consumption among those with the lowest educational level.
Data availability statement
No data are available. Not applicable.
Ethics statements
Patient consent for publication
Ethics approval
This study involves human participants, and ethical approval for the study was provided by the legal department at the University of Southern Denmark. According to Danish law, individual-level linkage of data from, for example, administrative registries is allowed without further consent when it is for research purposes and when thoroughly ensuring that results are presented in an anonymised way.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Contributors HARJ: methodology, visualisation, formal analysis, writing—original draft preparation. SRS: methodology, visualisation, formal analysis, writing—reviewing and editing. AIC: conceptualisation, methodology, writing—reviewing and editing. MD: conceptualisation, methodology, data curation, formal analysis, validation, writing—reviewing and editing. KJ: conceptualisation, methodology, data curation, formal analysis, validation, writing—reviewing and editing. CBP: conceptualisation, methodology, writing—reviewing and editing, guarantor.
Funding The present study was funded by the Ministry of the Interior and Health of Denmark. Grant number is not applicable as the study is part of the public sector tasks for the Ministry, which are governed in a framework agreement.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.