The scope and challenges of epidemiology

Table of Contents

The scope and challenges of epidemiology
Initial notes
Annotations on common readings
Annotated additions by students
Idea 1: The uses of epidemiology are many, but shift over time, and are subject to recurrent challenges from inside and outside the field.
Idea 2: In advising on the most effective measures to be taken to improve the health of a population, epidemiologists may focus on different determinants of the disease than a doctor would when faced with sick or high-risk individuals.
Idea 3: Epidemics affect population health, but are not the primary focus of social epidemiology.

Initial notes

On Idea 1
The articles provide a variety of historical perspectives and opinion statements on idea 1.
On Davey-Smith (2001), see Davey-Smith’s conversation with Jerry Morris, the author of the Uses of Epidemiology, http://bit.ly/g9DDHz
Brandt and Gardner's title conveys the point: physicians have often opposed an increasing role of public health. Epidemiology might be needed for quantitative assessment of new interventions and evaluating patient safety and healthcare quality, but its role beyond evaluation and assessment, especially in regards to social, cultural, and economic factors of diseases, is contested.
Pearce argues that modern epidemiologists have little concern for the socioeconomic factors that may affect health. He contrasts “bottom-up” and “top-down” approaches. The latter begins at the population level in order to determine the primary factors that effect health, and it uses a structural model of causation. The bottom-up approach, e.g., molecular epidemiology, begins on the individual level and aims to proceed upward toward the population level.
Putnam elaborates on Pearce's argument by discussing the macro-determinants of health. Krieger 2014b, as well as described in McMichael's review, provides a unified, politically engaged, multi-level view of the scope of epidemiology.

On Idea 2
Rose (1985) promotes the population health focus, but this is not universally accepted by healthcare practitioners and policy makers. If someone asks you the question Rose’s mentor posed, “Why did this patient get this disease at this time?,” how do you answer? Can you identify areas in your own life and/or work when you would take a population view and other areas where your focus would be individually-centered?

Road accidents and alcohol consumption may be a good illustration of Rose’s argument. Most of us know of getting home safely when we’ve drunk too much “risk factor,” but we also know that a substantial fraction of people in accidents have high alcohol levels. We also sense that some people are more susceptible to having their judgement and reaction times impaired by alcohol so we could imagine doing further epidemiological and biological research to develop multivariable risk factor formulas. Would a more refined knowledge of riskiness help us prioritize our risk-prevention efforts, or would that pale into insignificance relative to a Rosean drink-don’t-drive efforts?

Controversy over vaccination of girls for HPV, given the physical side effects (at a low rate — see http://www.usatoday.com/news/health/2009-08-31-hpv-gardasil_N.htm) and promiscuity-inducing side effects (no data for this). Question: What would Rose propose?

Question: Why isn’t a population an aggregation of individuals and thus population risk = sum of individual risks?
My response: 1. It is necessary to think of different meaning of “treatment.” A sick individual is treated by a physician to cure or reduce the effects of the disease. Population health policies do not treat a large group of sick people, but attempt to reduce the incidence in the next generation.

2. A physician treating sick individuals adjusts the treatment for individuals if it doesn’t work well for them. In contrast, public health measures usually discount the heterogeneity in the population and apply the same policy to all. Nevertheless, it is possible to imagine that knowledge of heterogeneous responses to treatment of individuals could lead to more effective population health policies (and reduce the kickback that occurs when some individuals claim to have suffered under the population health policy).

Davison (1991) addresses the individual-population contrast as it plays out in in "lay" (actually lower SES) views of their own health risks. (We return to this reading in the last session.)

On Idea 3
tba Hoffman....

Mini-lecture
Notes and annotations from 2007 course, 2009, 2011, 2013-
Common readings: Davey-Smith 2001 (uses of epidemiology), Rose 1985 (population health)
Supplementary Reading: Brandt 2000, Caldwell 2001, Davison 1991, Hoffman 2014, Krieger 2010a, McMichael 2011, Nandi 2014, Pearce 1996, Putnam 2008, Schwartz 1999

Annotations on common readings

George Davey Smith. The uses of 'Uses of Epidemiology'. International Journal of Epidemiology (2001) 30 (5): 1146-1155
In this article, George Davey Smith seeks to review the revolutionary impact of the text, ‘Uses of Epidemiology’ written by Jerry Morris, first edition published in 1957. Smith credits the text’s methodical approach for helping to actually create the field it sort to document, the population aspects of non-communicable disease. Smith’s primary conjecture is that much of the prescient ideas contained in ‘Uses of Epidemiology’ have been established as the cornerstone of contemporary epidemiology. However, the article acknowledges a trending departure from Morris’s school of thought asserting population-based modeling as being a best practice in uncovering determinants of diseases; the consequence of which is inferred to be adverse. (SY)

Rose G (Department of Epidemiology, London School of Hygiene and Tropical Medicine, Keppel Street, London WC1E 7HT, UK). Sick individuals and sick populations. International Journal of Epidemiology 1985;14:32–38

In this article, Rose compares the two approaches applied to aetiology, individual and population-based, and subsequently explores their counterparts in prevention. Determinants of individual cases and the determinants of incidence rate are distinct in nature, thus they require differing studies to elucidate causal agents depending on the reference frame, is the primary premise of this article. Rose examines the limiting effect of the underlying assumption of heterogeneity of exposure within the study population that is inherent to case control and cohort methods. Furthermore, Rose discusses the tendency for genetic factors to dominate individual susceptibility yet explain little of populational differences in incidence. Rose argues, in the unlikely event that the individual and population-based approaches to aetiology are in competition, preference should reside in discovering and controlling the causes of incidence; as susceptibly, indelible in its constitution, ceases to matter in the absence of determinants. (SY)



Annotated additions by students

(In alphabetical order by author's name with contributor's initials and date at the end.)