Life course epidemiology

Idea: How do we identify and disentangle the biological and social factors that build on each other over the life course from gestation through to old age?

Initial notes on the Cases

From PT (rev. 11/2):
The readings fall into three groups: Fetal & developmental origins of diseases in late life (Barker being generalized by Ben-Shlomo); consideration of cohort effect & general reviews by Lynch and Davey-Smith with Berney reviewing health in old age; and Brown on life course influences on depression (not necessarily in old age). Try to get a handle on all three groups before reading every page of any one group. Also, think about how Brown's method contrast with the bio-social epidemiologists'. In what ways could either side usefully draw methods, data, results from the other?

Substantive statement

We are going to compare and contrast the data, analyses, graphs, and figures to see how authors/researchers construct the causes built up over the lifetimes.

~~We will talk about how authors gather those data to make the schema and the life course perspective- kaorim07 kaorim07

a. Figure 3 Lung Cancer and Cigarette Consumption (Lynch & Davey
-Smith, 2005);
b. Infant Mortality and Stomach Cancer Mortality (Figure 2 on Lynch & Davey-
Smith, 2005; Figure 10 on Davey-Smith, 2007)
c. Figure 1 of Ben-Shlomo (2002) and Figure 1 of Brown (1978)

-When is a good point to gather data? and how?
-Whom do we want them to gather data and where should we store the
information?
My thought: General health information may be provided if both father or mother, or either one of them provide their consent. At any point after high school years (or even during the high school years), they can decide whether they want to match their more detailed information with their birth information at a place such as DMV. A small survey can be distributed at DMV also, since there is always a waiting line. A longer survey may be given at universities, work places, and by health care providers. Every five year may be desirable, and also give each respondent an option to participate in interviews for qualitative data collection. To have good cohort information, any kind of incentives may be helpful, such as giving $100 after every 3-decade (6 surveys). Information gathered can be stored at local level (e.g. universities) and share it within the state and with other states.

My thoughts: It is important to measure and treat depression. It is an outcome of a continuum of an individual’s life experiences. Besides chemical imbalance in one’s body, other psychological and social factors affect one’s mind. How resilient one is from experiencing a series of life events is hard to compare with another one. The degree of symptoms need to be treated may be differ. However we can’t ignore the golden standard. We may need to have a better understanding of untreated individuals, and those who are close to be treated. Does a collective culture have more pressure on individual? Would it be easier to spit out a negative experience first and consume a positive experience? Some individuals may be just receiving a free therapy from their social network over the lifetime. It is a question of how great a life event (e.g. having a great husband) needs to be to compensate a negative life event (e.g. mother’s death).

structural equation modeling for causation analysis in the following week 10.




*Afer Our Class Meeting*

“A life course approach too chronic disease epidemiology uses a multidisciplinary framework to understand the importance of time and timing in associations between exposures and outcomes at the individual and population levels.”
By looking at the Figure 2 Infant mortality rate and stomach cancer mortality rate for men aged 65 to 74, and Figure 3 Cigarette consumption and lung cancer (Lynch & Davey Smith, 2005), they make us think what makes these three countries alike, and if all smokers are going to suffer from lung cancer 30 years later. It is probably how people start to speculate the development of chronic disease and tease out a possible cause.

If we just look at two points at life time, birthweight or just health in childhood, and a chronic disease at some point in adult life, we could see the association between them as Baker, and Berney suggest. Our physiology is developed before we are born. Therefore, any dysfunction or low birthweight would have some influences on later life. Ben-Shlomo and Kuh (2002) and Davey-Smith (2007) are then aware of the importance of taking into account of gestation besides looking at other factors in the life course.

Ben-Shlomo and Kuh (2002) use a scheme to explain lung cancer by using four pathways (biological, sociological, socio-biological, and bio-social pathways) where they include socioeconomic status, (psychosocial) behaviors, and environment factors besides utero development.

Similar to Ben-Shlomo and Kuh (2002), Davey-Smith (2007) thinks that socioeconomic status, which is inevitably tied to the history of the country, is the key factor to health inequality. But a selected health outcome and income may not show a significant relationship as they are within the country when comparing with other countries. He shows that although the inequality of socioeconomic factors has been observed in the U.S., the gap has become smaller in last decades.
By adding socio-economic disadvantages (differences) into the early-life origin of adult disease, the significance of the early life and adult disease disappears. It is showing how important it is to consider socio-economic factors on the top of infant health. Besides identifying the socioeconomic status, Davey-Smith (2007) also explicitly talks about psychosocial factors to add into the life course perspective, and is aware of the complexity of the influence of psychology to health.


Positive and negative life events could be great stresses for people. The meaning of an event differs from one person to another; therefore, how strongly one event affects one person different from the other is hard to measure. And there are so many elements (e.g. social support) associated with the development of psychological disease, such as depression. Brown (1978) shows in the Figure 1 that the available resources and the cumulative disadvantages and advantages could result in different psychological outcomes. Causation identified in the diagram may be difficult to conclude and apply to everyone since self-esteem and vulnerability may go back and forth as well as personality were not included, as Brown notes. But it is still a great segment to understanding the development of depression.

As the articles let us see, including many aspects (biological, social, and psychological factors) of the life course can help researchers to pin point the causality of chronic disease in a more accurate way than underestimate them. More research using data from several cohorts, maybe along with twin studies, and a longitudinal analysis is needed. (km)




Response


The subject of today’s class is the life course approach to epidemiology. When presented as a topic for discussion during one class period with a set of readings assigned to this approach, it is easy to imagine that the life course perspective is discretely defined with a set of methodological assumptions agreed to by all researchers engaged in taking this approach. However, the period of the life course is of course extensive, and while the life course perspective does involve certain foci and underpinnings that may not be common to other strands of epidemiologic research, it nevertheless encompasses such a broad spectrum of possibilities for disease etiology and manifestation that it deserves intensive examination from various points of view.

In their assigned article for today’s class, for example, Ben-Shlomo and Kuh define the life course approach “as the study of long-term effects on chronic disease risk of physical and social exposures during gestation, childhood, adolescence, young adulthood and later adult life. It includes studies of the biological, behavioral and psychosocial pathways that operate across an individual’s life course, as well as across generations, to influence disease.” Berney et al. utilize a slightly different definition: “the ‘lifecourse perspective’ holds that inequalities in the structure of society shape life chances so that advantages and disadvantages cluster cross-sectionally and accumulate longitudinally.” What is immediately interesting about these two definitions is that, though they include similar elements, their emphases differ. Whereas the latter definition hones in on social inequalities as the primary driver of later life disease outcomes, the former definition takes a more cautious attitude and attributes chronic illness to more general “physical and social exposures.” Additionally, Ben-Shlomo and Kuh explicitly mention gestation as a time period in which these exposures may inflict long-term damage, thereby incorporating the fetal origins hypothesis (a.k.a. the Barker hypothesis) into their definition.

This raises the question of where the fetal origins hypothesis fits in epidemiologic analysis – is it a fully-formed approach unto itself, or must it be combined with more cumulative risk approaches to fully explain health outcomes in later life? However, this is not a question solely reserved for the fetal origins hypothesis; it applies to many other areas of literature in epidemiology as well. For example, Ben-Shlomo and Kuh also mention the study of allostatic load in their article, which calls up another broad literature that tries to account for the manner in which physiologic processes are disrupted by the social, environmental, genetic and geographical factors such as those referenced in the life course definitions. Many of the same elements are involved in this literature, but the methodological approaches and the objective focus of many of the studies differ. This week’s readings also include studies of mental health and depressive outcomes, which are often neglected in the health disparities literature and discussions of the impacts of cumulative risk and the influence of fetal origins, except where social status acts as a proxy for mental health (e.g. control over environment, social integration, etc.)

These examples of the types of epidemiologic research underway merely scrape the tip of the iceberg, though they do represent some of the more salient and dynamic approaches in the field currently. By taking a broad brush to the problem of chronic disease, the life course approach tries to envelop all the categories of factors that may play an influential role in determining adult health outcomes. The fetal origins hypothesis, taken as part of the life course approach or as an approach by itself, also adds an emphasis on the significance of geography. Again, taken together, these approaches seem to include a dizzying variety of factors related to health outcomes. Even so, as Ben-Shlomo and Kuh, and Barker (2007) indicate, these approaches offer a new way forward in epidemiologic inquiry. The life course approach attempts to fill in crucial gaps that other research has so far failed to account for and answer for the less than satisfying results that a focus on “the adult lifestyle model of chronic disease” has so far yielded. (291) In this regard, the life course approach does indeed seem quite promising and exciting. Yet, this does not mean that the life course (and fetal origins theory) is not fraught with obstacles related to specificity, measurement, categorization, application and so on. As Ben-Shlomo and Kuh state, “it remains to be seen whether as researchers we can cope with all this complexity.” (291)

Thus the incorporation of so many complex constructs into one epidemiologic approach begins to raise questions that it will be the task of epidemiologists and those interested in public health to attempt to resolve. Is it advisable to draw lines, and if so, where do we begin to do so? How do we achieve transdisciplinary study and perspective if the way to manage such a broad approach is to cut it into pieces according to expertise or discipline? In this way the life course approach may mirror the field of epidemiology in general – how do we make proper determinations about where to focus research efforts and resources in order to deliver the greatest benefit to public health? To establish its full contributions to the field of epidemiology, future research using the life course perspective will need to begin to examine the implications of these questions and many others while simultaneously providing nuance or clarification to the overall theoretical construct. (lh)



Annotated additions by students


Bengtsson, Tommy & Martin Lindstrom (2000) Childhood misery and disease in later life: the effects on mortality in old age of hazards experienced in early, southern Sweden, 1760-1894; Population Studies, Vol. 54, No.3. , pp. 263-277.
The aim of this study was to investigate cause-specific mortality at ages 55-80 years in four parishes in Scania in the southernmost part of Sweden during the period of 1760-1894. The authors interrupt old-age mortality in 2 ways.
  1. as caused by events in early life which lead to irreversible or partly irreversible tissue damage but with a long interim without symptoms.
  2. as the result of stressful life events in later life that were the proximate causes of death
They draw on the D.J.P.Barker hypotheses that a relative lack of nutrition before birth causes irreversible tissue damage and might even disrupt hormonal and metabolic pathways causing cardiovascular and diabetic disease later in life. Nutritional status was measured by the price of rye (the dietary staple) during the years proceeding and following birth and the year before death. Findings were that the disease load during the first year of life is the early-life indicator variable that significantly influences mortality in later life. Noted was the adverse impact on the development of the child of a high-mortality environment during the first year of life has an enduring impact on health and mortality in later life. No such influence on old age is found for the early-life variables that indicate stressful events during inter-uterine life (such as the relative lack of nutrition) during the early childhood period from 0-5 years. It is important to note that although the Barker hypothesis did not hold up, these results are inconclusive concerning it. More valid measures of nutritional status during pregnancy and early life are needed plus accurate diagnoses of what people actually died from should be offered. This was an interesting historical study but more timely research is needed. It would be interesting to see this study replicated during the 20th century.(jg)

McEwen BS. Early life influences on life-long patterns of behavior and health. Ment Retard Dev Disabil Res Rev 2003;9:149-54.

In this article, Bruce McEwen who is one of the more instrumental researchers involved with the conception and analysis of allostatic load (“cost or wear and tear on the body produced by repeated activation of the stress-responsive biological mediators such as glucocorticoids and catecholamines”), outlines possible physiological mechanisms through which lack of stability in early life may affect long-term mental health outcomes. Specifically McEwen focuses on how inconsistent care-giving, neglect, and abuse in childhood may alter brain chemistry and composition such that people become more prone to mood disorders and heightened stress responses later in life, which in turn may lead to increased morbidity and mortality from a variety of factors. He discusses how the amygdala and the hippocampus may be especially compromised as a result of adverse events in early life, hindering the developmental adaptation that would otherwise serve to regulate physiognomy as children grow. McEwen concludes however, that intervention and prevention are possible, and more than that, there is evidence to support the idea that such changes to brain chemistry are not entirely irreversible – a topic for future research. (lh)

Krieger N, Chen JT, Coull BA, Selby JV (2005) Lifetime socioeconomic position and twins' health: An analysis of 308 pairs of United States women twins. PLoS Med 2(7): e162.

This study compared the health status of 308 women pairs of monozygotic and dizygotic twins from the Kaiser Permanente Women Twins Study Examination II who had lived together until the age of 14 and subsequently apart into adult life. The authors hypothesized that "cumulative experiences across the lifecourse, including those after adolescence and after completion of educational attainment, and not just early life experiences, shape adult health." They looked at how much education each twin had and their socioeconomic positions in later life, and they analyzed these in relation to diverse health outcomes (blood pressure, cholesterol, body mass index). Findings were that while little differences were observed in health outcomes as related to educational attainment, monozygotic twins who differed later in life in their social class (categorized as 'working class' or 'nonworking class') did have differences in health, with the working-class twin having higher blood pressure and higher cholesterol than her nonworking class sister. The section on the weaknesses and strengths of the study was helpful, as was the detailed description of the methodology and measures used, and the specifics of the data analysis. [PLoS Medicine is an open-access journal which allows anyone without access to expensive personal or library subscription databases to find and read scientific literature. That more scientists and scholars are utilizing open-access journals - in addition to the 'big guns' in their disciplines - is a good thing, IMHO.] (JC)

Barker DJP. The origins of the developmental origins theory. Journal of Internal Medicine 2007;261:412-7.

In this recent issue of the Journal of Internal Medicine, DJP Barker and several of his colleagues publish articles on the origins of the Barker hypothesis and results from subsequent research in the area. This article gives Barker the opportunity to discuss the manner in which developmental origins theory evolved. Expanding on the explanation he provides in the first chapter of his most recent book, which we read for an earlier class, in this article Barker describes why behavioral factors are not a sufficient means of characterizing the changes in the incidence of heart disease in many industrialized countries in the past century. He also further expounds on the importance of geographical study in revealing the relationship between neonatal environment and heart disease. Barker then itemizes some of the more recent studies conducted in countries other than the UK (Finland, U.S., etc.) that add credibility and context to the developmental origins theory and suggests that the theory “offers a new way forward” when attempting to understand chronic disease. (lh)

Patterson Paul H. "Maternal Effects on Schizophrenia Risk"
Science 26 October 2007 318: 576-577
"Local environmental factors impinging on the fetal brain can tip the balance toward
mental illness." A different kind of fetal origin of an adult condition.