Background
Lung cancer (LC) is one of the most common cancer diseases and ranks among the most frequent causes of death in Germany [
1,
2]. With regard to the development of incidence over time, different trends were observed for men and women. In the last decades, a decreasing number of incident cases was reported for men. In contrast, the yearly incidence in women rose continuously [
2‐
4]. The strong link between socioeconomic status (SES) and morbidity has been emphasised in many studies (e.g. for Germany [
5,
6]). International studies show that there are also strong inequalities with regard to LC, especially among men [
7‐
16]. So far, however, little is known on how social inequalities in the incidence of LC developed over time. This holds especially true for Germany.
Few international studies investigated trends in social inequalities in LC morbidity and mortality. These studies mostly reported increasing social inequalities for men and women [
11,
12,
15,
16]. Studies investigating social inequalities in the incidence of LC among the German population are rare. This is due to the fact that the data required for such analyses are scarce as official cancer registries do not include socioeconomic information of the diseased individuals. Current research from Germany based on cross-sectional data investigates whether social deprivation, measured by regional levels of income, education, and labour force participation, is associated with cancer. For these analyses, the macro-level information on regional social deprivation on district level was combined with cancer site-specific incidence rates. For LC, a social gradient was found in men but not in women [
9]. However, as this solution is susceptible to ecological fallacies, the use of SES information on individual level should be preferred whenever possible. Furthermore, studies analysing time trends in SES inequalities in LC incidence in the German population are still lacking. Our study aims to step into this gap by using individual data on SES and incidence to examine time trends in inequalities in LC in Germany.
From a life course perspective, the cumulative disadvantaging effects of potentially harmful behaviours determine the outcome of social inequalities in health and mortality in later life [
5,
17‐
19]. This holds especially for the harmful effects of smoking, as the risk of developing smoking-related diseases increases with the duration of smoking. The earlier individuals start smoking the more likely they are to suffer from a smoking-related disease later in life [
20,
21]. Until the 1970s, smoking became more and more common but increasingly restrictive tobacco control policies in the 1990s and later had led to declining smoking rates, especially among highly educated individuals [
17,
18,
21‐
26]. This development also had a positive impact on trends in mortality and life expectancy [
27]. At the same time, smoking rates among individuals with low SES were quite stable or decreased at a slower pace than the rates among highly educated individuals [
21,
26]. Moreover, a convergence of smoking rates between men and women has been observed in many industrialised countries since the 1960s and 1970s [
18,
21,
25]. This convergence is rooted in changes in the social position of women and the adoption of risky health-related behaviours [
18,
21,
25,
28,
29]. This has led to declining differences in smoking-related morbidity between men and women [
13,
18,
21,
25,
28‐
32]. As a result, the burden of smoking-related diseases can be expected to concentrate in groups with lower SES. However, since the trends in smoking vary between sex and age groups, differing trends in social inequalities in LC incidence between men and women may also be expected [
21,
33].
Due to data restrictions, many studies combined different indicators of SES (e.g. income, educational level, and occupation) into a single deprivation index. However, previous research has shown that these indicators measure different aspects of social inequality and should therefore be analysed independently whenever possible [
34].
The aim of the study is to investigate time trends in LC incidence in Germany. Special attention will be paid to time trends of social inequalities in incidence and whether these trends differ between men and women, and between younger and older age groups. The analyses are based on claims data of a large German statutory health insurer, which contains large case numbers and different information on SES characteristics of the insured individuals.
The study is guided by the following research questions:
1.
Are there socioeconomic inequalities in lung cancer incidence? Do these inequalities exist in men and women, and all age groups equally?
2.
Are there different time trends in lung cancer incidence between socioeconomic groups? Are these trends in inequalities similar in men and in women?
Discussion
The aim of the study was to investigate time trends in socioeconomic inequalities in LC incidence in Germany. Our results concerning time trends in sex-specific incidence of LC are in line with the official statistics for Germany [
2]. We found decreasing incidence rates in LC for men and increasing rates for women. This development had an impact on the gap between the sexes as the differences were reduced over time. In accordance with previous research [
7‐
16], our findings reveal higher social disparities in LC for men than for women. These disparities were most pronounced in terms of income inequalities in LC incidence in men. Driven by the decline in LC incidence among the middle- and higher-income group, income inequalities widened in men. This increase in inequalities was strongest in men at retirement-age. In contrast, a reversed income gradient was found among women in the period 2006–2009. Due to rising LC incidence rates among women with low incomes and the decline in rates among women at working age with higher incomes, this gradient disappeared over time. Among men at working-age, the strongest inequalities in LC incidence were found in terms of income, the weakest in terms of occupational group, while among women, educational inequalities were strongest. Overall, occupational and educational inequalities tend to narrow over time.
Our study is one of the few that examines time trends in socioeconomic inequalities in LC incidence [
11,
12]. International studies reported increasing inequalities in terms of social deprivation levels in LC morbidity for men and women for Great Britain [
12,
16] and New Zealand [
11]. Our findings are in line with current findings on LC incidence by regional deprivation level in Germany based on cross-sectional data, which reported substantial inequalities in men but not in women without considering differences in the working-age and retirement-age population [
9]. In our analyses we may have indirectly depicted the results of changing regimes in harmful smoking behaviour in the 1960s and 1970s [
18,
21,
25,
33], especially in women with higher SES, which is in accordance with a recently published study [
39]. This underlines the importance of analysing time trends in SES inequalities in LC since cross-sectional approaches do not allow to depict the dynamics in cancer incidence within the different SES groups over time. This holds especially true for women. Thus, our findings may depict the effect of increased efforts in smoking prevention of the last decades [
22,
24]. It should be noted that established tobacco prevention measures in Germany are weaker than in other European countries. Although progress has been made (e.g. smoking ban in public places), stricter anti-smoking measures are often called for but have not yet been implemented [
4,
44]. However, our findings of the present and a recent study [
39] indicate that 20 to 30 years later previous efforts in smoking prevention seem to have an impact on LC rates in women and men. Nevertheless, smoking-related diseases are still among the main driving forces of sex differences in mortality in Germany [
32]. In contrast to the development at retirement age, inequalities among the working-age population are persisting or tend to narrow over time. This holds for educational as well as for occupational inequalities. With respect to occupational inequalities, this can most probably be explained by increasing standards of occupational safety, which among other things, resulted in lower exposure of carcinogens in the working environment over time. Among women at working-age, inequalities were strongest in terms of education while income inequalities in LC were much weaker than in men. This may be explained by the overall lower income level among women, which persists even when men and women with the same level of education are compared [
45].
The findings indicate that trends in social inequalities in LC incidence may vary between age groups or birth cohorts. This holds especially true for women. However, as the majority of incident cases lies above the age of 70, the analyses concerning educational level and occupational positions are based on low numbers of incident cases. Therefore, it is important to continue to monitor developments to see whether trends in educational and occupational inequalities continue into old ages.
Strengths and limitations
Our study is based on health insurance data spanning the time period from 2005 to 2017 that provide high case numbers and includes information on individual diagnoses as well as on socioeconomic characteristics. This permitted to analyse time trends in social inequalities in LC based on three SES indicators and among different age groups.
A major strength of our data is that all information is available at individual level, which prevents economic fallacies in the interpretation of the results [
39]. Furthermore, the data contain the complete insurance population and are therefore not subject to selection bias with regard to health status [
37‐
40,
46]. More detailed information on general strengths can be found in previous studies (e.g. [
37‐
40]).
The precise information on individual income and high case numbers allow to obtain a detailed picture of the development in LC incidence in different income groups from age 20 to the oldest old as well as in the population subgroups at working and at retirement age. As for other studies based on health insurance data the results concerning income inequalities should be interpreted carefully since the data do not include information on household income (e.g. [
37‐
40]). However, previous analyses have shown that social gradients in health obtained from estimates based on household income can largely be replicated using individual income, which suggests individual income to be an adequate measure to study social inequalities in health [
47].
The data on educational graduation and occupational position are restricted to the population at working age. Due to this limitation, the analyses of trends in educational and occupational inequalities could not be conducted for individuals above age 65. It can be assumed that inequalities in LC caused by the former occupational position or the educational level persist into old age. Additional information on educational level and former occupation would have allowed a deeper insight into the developments of LC inequalities, especially among women, but cannot be analysed on the basis of our data.
As described in previous studies based on this data (e.g. [
37‐
40], the data are representative for the total population of Germany in terms of sex and age structure but differ in terms of social distribution [
36]. We addressed this limitation by stratifying or controlling all analyses for socioeconomic indicators. Therefore, the reported results should be unaffected.
Conclusion
Our study reveals that social inequalities in LC are considerable and that trends vary with respect to SES group. The findings indicate that income inequalities widened among the elderly, but occupational and educational differences remained fairly stable or even narrowed among the working-age population. Most disadvantaged are men at retirement age with low income, for whom the increase in inequalities was most pronounced. More research is needed to uncover the underlying mechanisms that explain the widening inequalities in men and the observed trends in women. Our findings indicate that time trends in LC differ not only with respect to SES but also according to age range or birth cohort. The findings also suggest that focussing on social inequalities without considering differences between age-groups and time trends in health inequalities could lead to existing inequalities remaining undetected. In LC, this holds especially for the trajectories in inequalities among women over time.
It is important to foster public health interventions (e.g. complete ban of tobacco advertising and smoking in public places) to reduce LC incidence in the German population. Against the backdrop of existing inequalities, interventions should mainly focus on deprived social groups.
Acknowledgements
We thank the AOK Niedersachsen (Statutory Local Health Insurance of Lower Saxony) for providing the data. In particular, the support of Dr. Jürgen Peter, Dr. Jona Stahmeyer, and Dr. Sveja Eberhard made it possible to carry out this study.
Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit
http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (
http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.