Life course researchers are particularly interested in the “social patterns in the timing, duration, spacing and order of events and roles” (Elder and Rockwell 1979, p. 2). Attention is paid to how certain transitions or events can produce different effects depending on their timing within the life course (George 1993). For example, the consequences of the Great Depression were different for older and younger children (Elder 1974). In addition, in life course epidemiology the notions of timing and duration are central to the three main models for the association between early life circumstances and later life: the latency, pathway, and accumulation models (Graham 2002).

By contrast, the temporal dimension of preventive behaviors has been generally ignored in both empirical research (Spadea et al. 2010) and medical sociological theory (Missinne et al. 2014). This is somewhat unfortunate, as the effectiveness of care depends upon a timely initiation of preventive care or check-ups and upon its regular use. As a result of the focus on rates of illness-related health care use in the Andersen’s framework (1983) and the dominant use of cross-sectional study designs, questions about (preventive) health care use are formulated along the lines of: “during the last xx months/years, have you consulted a specialist/GP/dentist/had a mammogram ?”. This design and this question wording render it impossible to scrutinize both the timeliness and regularity of preventive behavior. To capture a regular pattern of care, the perception of a ‘usual source of care’ (e.g. “is there a particular doctor you usually go to when ill, or for advice about health?”) is also often used. However, this type of measurement also fails to adequately capture periodic behavior and even the preventive nature of a visit (Newman and Gift 1992).

Timely detection of breast cancer is crucial given that the stage of illness (or tumor size) at diagnosis is strongly linked to survival (Elmore et al. 2005). Therefore, the Council of the European Union recommends that screening programs target women aged 50 to 69 years of age (von Karsa et al. 2008), who are at the highest risk of breast cancer. Age is generally regarded as a control or a confounding variable, or is used as a proxy for ‘need’ for general health care use (Van der Heyden et al. 2003), preventive health care use (e.g. Jusot et al. 2011), and for mammography screening (e.g. Duport and Ancelle-Park 2006; Wübker 2012). In addition, the regularity of preventive habits is recommended. For example, a two-year interval is recommended for mammography screening (European Commission 2003), six months for dental check-ups (Riley et al. 2013) and targeted groups should be given a flu vaccination every year.

In the discussion concerning the social gradient, the temporal dimension should also be included. It is possible that socioeconomic inequalities are partly manifested in both the regularity and the timeliness of preventive health care use, in addition to the probability of ever engaging in it (Missinne et al. 2014). For example, health insurance data has shown that in Belgium, the dropout rate for two-yearly mammography screening is higher among women who benefit from preferential reimbursement (Fabri et al. 2010). Higher-educated groups might be more future oriented and more willing to commit to a long-term goal, such as prevention (Mirowsky and Ross 2003; Wübker 2012). Accordingly, it is possible that cultural health capital includes knowledge or competencies that enable the timely and regular use of preventive health care.

The way the SHARELIFE investigates mammography screening includes both notions of temporality and regularity. The question, “In which year did you start having mammograms regularly?” was given to all women who answered yes to the question “Have you ever had mammograms regularly over the course of several years?” This retrospective information allows us to gain insight into two important questions relating to the timeliness of mammography screening. The first question relates to the age differences that are often reported for mammography screening. Empirical studies generally report lower engagement in screening among older women (Wübker 2012), but confusion remains substantial (Jepson et al. 2000). The dominant question wording renders it impossible to know whether age differences reflect ‘true’ age effects or whether they act as proxies for period effects.⁴ The latter is very probable, as knowledge about – and policy initiatives concerning – breast cancer and mammography screening have changed considerably in Europe during recent decades (Fisher et al. 2008). To this end, an explorative approach has been followed. Five birth cohorts from 1910 to after 1949 in ten-year intervals were constructed and the Kaplan-Meier graphs were tabulated for the 13 European countries that took part in the SHARELIFE (see Fig. 6.2 for three examples; for details see Missinne and Bracke 2014). The results do indeed suggest substantial period effects. In all countries, earlier birth cohorts overall engage less in screening. Figure 6.2 show that the hazard function for each earlier birth cohort is lower at all ages, except for Sweden. Very similar age trajectories can be observed for each cohort, suggesting no ‘true’ age effects. The cross-national comparative approach aids in framing these period effects within the context of national screening policies, which have already been empirically linked to the large country-differences in mammography screening (Wübker 2014), illustrated in Fig. 6.4. I will return to this point when elaborating on principle 4. In addition, country-specific deviations can be related to features of national screening policies, again suggesting strong period effects. For example, the coinciding hazards of the two most recent cohorts, as well as an additional increase at the age of 40, can be traced back to the early implementation of a national screening program in 1986, which targets women from the age of 40 in 65 % of Swedish counties. In Belgium, women are invited to participate from the age of 50. In Greece, the absence of a sudden increase indeed reflects the absence of a national program (for an overview of screening policies, see von Karsa et al. 2008).

Next, the hypothesis that socioeconomic inequalities could also be manifested in the timeliness of the take-up of mammography screening was tested. Returning to Fig. 6.1 and focusing on the specific age trajectories, an increase of screening around the recommended age of 50 is notable for all social groups. These age trajectories do not differ according to socioeconomic indicators for either childhood (Fig. 6.1) or adulthood (Fig. 6.1). Accordingly, in line with what studies have traditionally assumed, socioeconomic inequalities seem to be manifested in Belgium as a lower probability of ever having a mammogram, rather than in the late commencement of screening. This finding should be interpreted in the light of the relatively small age range for which screening is recommended. The discussion about timeliness should therefore not be closed. For example, for preventive services that begin far more early in life, such as dental check-ups, timeliness might reveal clearer socioeconomic inequalities in preventive health care.

The principle of agency stresses that individuals are not passive recipients. Encapsulated in life course research is the question of how the interplay between individual action and the social structure shapes the lives of individuals. Individuals act and make choices within the opportunities and constraints of their world (Elder et al. 2003). For example, Elder described how parents and children successfully adapted to the difficult circumstances during the Great Depression (Elder 1974, 1998).

In the introductory section of this chapter, I already touched on the structure-agency debate, which has also received considerable research attention in medical sociology (Abel and Frohlich 2012), most directly when studying health inequalities (Cockerham 2005). Recently, medical sociologists have endeavored to theorize the relative importance of agency and structure for health and health lifestyles (Williams 1995; Cockerham 2005, 2007; Abel and Frohlich 2012). It is acknowledged that “counterposing agency with structure is a misplaced and false dichotomy” (Dannefer and Daub 2009, p. 20). Instead, they can be recursive (Frohlich et al. 2001) and the question is the extent to which either one is dominant in a particular situation (Cockerham 2007). Cultural health capital theory focuses on the specific situation of health care interactions. In this way, the broader macro-structural level of the unequal distribution of resources is linked to micro-level practices (Abel 2008; Abel and Frohlich 2012). Shim (2010) highlighted that individuals are not passive recipients of cultural health capital strongly tied to social stratification.

Another way to gain insight into the structure-agency debate is by focusing on the different socialization contexts socially-mobile individuals are confronted with over their life course. Each social position largely determines the ‘life chances’ of individuals at that time and these positions constitute the structuring forces of ‘life choices’ (agency) on health lifestyles (Cockerham 2005). The weight that Bourdieu attributed to childhood experiences in the formation of the habitus, has often been critiqued (Daenekindt and Roose 2013). Social mobility research parallels this idea by addressing the multiple contexts of socialization, each with its own health-related practices. Socialization continues into adulthood, when individuals are confronted with new experiences (Ryder 1965) and other significant network members become important for health behaviors (Christakis and Fowler 2007), for example marital partners (see the principle of linked lives). Social mobility research can gain insights into the development of health lifestyles by scrutinizing the relative impact of the social position in childhood versus the prevailing social position.

Using the example of mammography screening has two advantages when studying the health behavior of socially-mobile individuals. First, it is only recommended from the age of 50 onwards (WHO 2013), when social mobility processes are likely to have been actualized. Therefore, this type of health behavior is not likely to affect the course of social mobility. In most studies, such a process of reversed causality cannot be ruled out and hampers causal interpretations of the effect of social mobility (Claussen et al. 2005). Second, it is very unlikely that mammography screening is related to the event and the accompanying stress of social mobility itself, as has been suggested for health-compromising behaviors such as alcohol use or dietary patterns (Karvonen et al. 1999).

Diagonal Reference Models (DRMs) were designed in particular to study the effects of social mobility and enable estimation of the relative impact of the social position of origin and the social position of destination. The screening behavior of socially-immobile individuals is taken as the reference points, as they represent the core of each social stratum. Therefore, their health-related behavior is considered characteristic for that social position. Consequently, the health behavior of socially-mobile individuals is modeled as a function of the characteristic behavior of immobile individuals from the social position of origin and of destination (for an outline of the empirical strategy, see Daenekindt and Roose 2013; for examples in health research, see Monden and de Graaf 2012). These models were applied to the Belgian sample of the SHARELIFE to test three hypotheses: (i) health behavior is, in line with Bourdieu, predominantly shaped by the primary socialization context: the social position of origin; (ii) the health behavior of socially mobile individuals is predominantly associated with the social position of destination; and (iii) the maximization hypothesis considers whether the experience of upward social mobility differs from that of downward social mobility. The results showed that the take-up of mammography screening by both upwardly and downwardly mobile individuals reflects the patterns of the women in their prevailing social position. Therefore, empirical support is only found for the second hypothesis. This points to the situational nature of mammography screening, which is also highlighted in the empirical example outlined in the next principle. However, it does not necessarily contradict our findings that childhood socioeconomic conditions are crucial in the development of cultural health capital (principle 1). The data limited us to considering only occupational mobility, which is regrettable given that the role of cultural capital in particular, such as education, has been seen as increasingly important.

With its principle of linked lives, the life course perspective highlights that individual lives are lived interdependently in a network of shared relationships (Elder et al. 2003). Because experiences are shared, the relevance of various social events and transitions is widened (Heinz and Kruger 2001). These interpersonal experiences are also located within a specific historical time and place that can impact upon these micro-level settings (Elder et al. 2003).

Research on preventive health care and health care in general has focused too much on the individual in isolation. Andersen’s (1970, 1995) heuristic model, focuses on how individual need, socioeconomic and demographic characteristics, and individual health beliefs are related to health services use. However, seeking professional care is often not the result of an individual decision, but of an interactive process (Pescosolido 1992). Recently, Umberson and colleagues (2010) drew explicitly on the life course perspective to provide a theoretical framework to unfold the mechanisms underlying the relationship between social ties and health behavior, including preventive health care use and treatment attendance. Predominantly, the focus is on the presumed beneficial effect of marriage. Health-related social control theories propose that partners try to influence and regulate each other’s health behavior in order to keep their partners healthy (Lewis et al. 2006; Umberson 1992). However, the universal protective nature of marriage has been challenged (Carr and Springer 2010). Again, the discussion is hindered by the wide use of cross-sectional designs, which make it impossible to discern to what extent the effects attributed to marriage can also be ascribed to premarital health habits and premarital socioeconomic conditions (Meyler et al. 2007). Individual lives are not unwritten pages at the time of marriage. As outlined in the life-span development principle, conditions earlier in life are crucial to the development of health behavior. Although marital partners are the most important and powerful source of influence in a person’s adult life, parents are predominant during childhood (Umberson 1992), and also influence socialization into healthy behaviors (Cardol et al. 2005).

In Missinne et al. (2013), we hypothesized that cumulative life course advantages or disadvantages accumulate at the household level and will be greater than at the individual level. Partners provide each other with information and norms on health behavior (Thomas 2011). Therefore, it can be expected that (un)favorable socioeconomic conditions for either partner in childhood will impact on health behavior in later life. Assortative mating can exacerbate these effects and generate systematic divergences over the life course, as contented by cumulative advantage theory (DiPrete and Eirich 2006). Cultural health capital theory might benefit from the explicit inclusion of the notion of linked lives. To elaborate on how capital is acquired and accumulates over time, it is important to understand the role of the childhood and adult preventive health behaviors of both partners.

The dyadic nature of the SHARE and SHARELIFE enabled scrutinizing the influence of the childhood preventive health care behavior of both wives and husbands on the initiation of mammography screening for a sample of Belgian women (N = 734). Figure 6.3 show the Kaplan-Meier graphs for the married women in our sample. As in Fig. 6.1, childhood preventive behavior seems to impact on mammography screening many years later in life (p < 0.001). However, what is new is the significant effect of the partner’s preventive behavior (p < 0.001). Again, the complementary log-log models show that the latter association persists after controlling for the characteristics of the women (number of books during childhood, childhood and lifelong dental check-ups, educational level), the characteristics of the men (childhood and lifelong dental check-ups, educational level), and household wealth. More precisely, the hazard for mammography screening in later life is 25 % higher for women whose husband went regularly for dental check-ups as a child. The results suggest that the cultural health capital of both partners impacts on women’s preventive health care use and show the importance of the contextualization of preventive health care use within the family.

This principle refers to the fact that the life of every individual is embedded in and shaped by a certain historical context and place (Elder et al. 2003). Historical change can impact on an individual’s life. It can also engender cohort effects when it alters the lives of successive birth cohorts, and period effects when the effect is more uniform across these cohorts (Elder 1994). The different aftermath of World War II in Europe than in the United States, illustrates that historical events might impact differently across regions or nations (Elder et al. 2003). To translate this idea into empirical practice, life course researchers urge us to expand the scope from national studies, the results of which can be challenged as being too context specific, into international comparative studies (Billari 2009; Blane et al. 2007). This has now become possible in Europe with the advent of some large-scale research projects that have collected life course data which is fully internationally comparative, such as the SHARE.

Although the cross-national comparative approach is well established in health (e.g. Mackenbach 2012) and health care research (e.g. Devaux 2013), it is still upcoming in preventive health care research (Jusot et al. 2011). However, the existing studies have already revealed substantial cross-national variation in preventive health care habits, including mammography screening (e.g. Wübker 2014). An important question now relates to which institutional differences are the driving forces behind this cross-national variation (Blane et al. 2007). For mammography screening, it seems that general (health care) indicators (such as health care expenditure, number of physicians, and gross domestic product) do not matter (Jusot et al. 2011), but that the country-specific characteristics of mammography screening policies should be focused upon (Wübker 2014).