0. introduction to epidemiology
- 1. Introduction PH250B Epidemiologic Methods II Fall 2013 Jennifer Ahern, PhD Fall 2014 Jack Colford, MD PhD
- 2. Introduction outline – Definition of epidemiology – Epidemiologic terms – Course in context • Questions and answers • Health paradigms • Specific or general? • Recurring issues – Summary
- 3. Epidemiology • What is epidemiology? – Study of the distribution and determinants of health- related states and events in populations • Includes physiologic states, psychological states, and positive outcomes – Clinical medicine also addresses health and disease in individuals, but the unique aspect of epidemiology is its focus on population distributions
- 5. Distribution of health • Distribution – breast cancer Incidence Rates by Race Race/Ethnicity Female All Races White Black Asian/Pacific Islander American Indian/Alaska Native Hispanic 127.8 per 100,000 women 132.5 per 100,000 women 118.3 per 100,000 women 89.0 per 100,000 women 69.8 per 100,000 women 89.3 per 100,000 women
- 6. Obesity Trends Among U.S. Adults 199 9 200 9 199 0 No Data <10% 10%–14% 15%–19% 20%–24% 25%–29% ≥30% Distribution of health
- 7. Determinants of health • World Health Organization • The determinants of health include: – the social and economic environment, – the physical environment, and – the person’s individual characteristics and behaviours
- 8. Determinants of health • Healthy people 2010
- 9. Determinants of health Results The leading causes of death in 2000 were tobacco (435000 deaths; 18.1%of total US deaths), poor diet and physical inactivity (400000 deaths; 16.6%), and alcohol consumption (85000 deaths; 3.5%). Other actual causes of death were microbial agents (75000), toxic agents (55000), motor vehicle crashes (43000), incidents involving firearms (29000), sexual behaviors (20000), and illicit use of drugs (17000).
- 10. What brings you here? • What “determinants” are you interested in? • What “health related states and events” are you interested in?
- 11. Introduction outline – Definition of epidemiology – Epidemiologic terms – Course in context • Questions and answers • Disease paradigms • Specific or general? • Recurring issues – Summary
- 12. Epidemiologic terms • “Determinants” of health also called – Exposures (used generally) – Treatment (used generally) – X (as in X and Y in a statistical model) – A (as in A and W and Y in a statistical model – A as exposure/treatment of interest, W as nuisance, Y as outcome) X Y
- 13. Epidemiologic terms • “Health related states and events” also called – Outcome (used generally) – Disease (used generally) – Y (as in X and Y in a statistical model) • Many terms seem to imply a yes/no categorization but they encompass any kind of health outcome (continuous, categorical, binary measures) X Y
- 14. Epidemiologic terms • Covariates – Other variables that have some influence on the relation between an exposure and outcome – X (“other” Xs) generically – W if the influence on the relation is a nuisance distortion you want to remove (aka, confounding) – Z if the influence on the relation is of interest (e.g., effect modification, mediation)
- 15. Epidemiologic terms • Strata – Subgroups of a population separated according to specified criteria such as age groups, sex etc. • Stratification – The process of separating a population into these groups
- 16. Epidemiologic terms • Associations distinguished from effects – An association tells us about probabilities of past events • Carrying matches is associated with lung cancer – An effect is causal and it tells us how probabilities change if conditions change • If you remove matches from pockets in the population, does the rate of lung cancer decrease?
- 17. Introduction outline – Definition of epidemiology – Epidemiologic terms – Course in context • Questions and answers • Health paradigms • Specific or general? • Recurring issues – Summary
- 18. Context
- 19. Questions and answers • Types of questions we can ask/answer with epidemiologic methods (Rose 1985) • Clinician’s question – “Why did this patient get this disease at this time?” • Causes of cases – “Why do some individuals have hypertension?” • Causes of incidence – “Why do some populations have so much hypertension whilst in others it is so rare?”
- 22. Questions and answers • Determinants of the mean are the causes of incidence • Determinants of location within the distribution are the causes of cases • Epidemiologic approaches can be used to identify either types of determinants – but are being applied in the service of fundamentally different questions • Depending on the question you ask you get a different answer
- 23. Questions and answers • What makes these fundamentally different questions? – If a population is just an group of individuals, how are the causes of cases and the causes of incidence different? – In a study focused on causes of cases you accept as constant background aspects the group(s) to which individuals belong that cause disease – In a study focused on causes of incidence you explicitly examine the effects of aspects of the group (s) to which individuals belong • Aspects of groups often driven by fundamental characteristics of a society/culture
- 24. Questions and answers • Example – causes of cases vs causes of incidence • Obesity – today’s big epidemic – what is the cause? – Epidemiologic studies of twins suggest 2/3 of obesity attributable to genetics
- 25. Questions and answers – US obesity prevalence has increased from 12% to 19% in less than 10 years – Are genetics behind this? – Must be caused by changing environmental factors
- 26. Questions and answers • Which is the right answer – genes or environment? • Both the right answers to different questions – Genes tell us about the causes of cases – Environment tells us about the causes of incidence • Where you focus your lens determines what you can see – What exposure and outcome studied – How much does the exposure vary across the population(s) studied over the time studied
- 27. Questions and answers P Menzel. Hungry Planet. What the World Eats. Ten Speed Press, 2005. Egypt USA Ecuador
- 28. Introduction outline – Definition of epidemiology – Epidemiologic terms – Course in context • Questions and answers • Health paradigms • Specific or general? • Recurring issues – Summary
- 29. Paradigms • Health paradigms shape epidemiology and epidemiology shapes them • Paradigms of health – Sanitary (18th, 19th c) – miasma – Infectious (20th c) – germ theory – Chronic diseases come to fore in developed countries • Multifactorial, lists of risk factors… • But wait… some of them are infectious?
- 30. Paradigms • Increases in peptic ulcer disease in first half 1900s was of great concern • Thought to be a “disease of civilization” • Susser and Stein documented birth cohort patterns suggesting peptic ulcer disease mortality had peaked and had started to decline as of 1960 – Met with great skepticism • Interpretation strongly influenced by current paradigm – chronic disease was non-infectious and dependent lifestyle in adulthood
- 31. Paradigms Mervyn Susser • “concluded that the apparent multifactorial aetiology of peptic ulcer – with contributions from diet, alcohol, cigarette smoking, emotional strain, personality and genotype did not ‘exclude the possibility that a major single causal factor awaits discovery’” (Davey Smith, Ebrahim 2001)
- 32. Paradigms • Heliobacter pylori infection was identified as the cause • Cause identified by a pathologist and clinician with no input from the extensive body of epidemiologic research – Defied a long held biological tenet – bacteria cannot survive in the acidic environment of the stomach
- 33. Paradigms • Went against the paradigm of chronic disease causality • Contradicted a biological tenet
- 34. Paradigms • Continuing discussion in epidemiology about whether health paradigms, conceptual frameworks etc. ground us or constrain us – Some argue epidemiology is at its best documenting patterns and associations without being influenced by existing paradigms and theories – Others say it is meaningless without theory
- 35. Paradigms • “Scientists rely heavily on unstated criteria, or degrees of confidence… in particular hypotheses or approaches, and investigators are only sometimes fully able or willing to acknowledge them… They include all of the vanities, vested interests, hunches, experiences, politics, careerism, grantsmanship tactics, competing cadres of collaborators, imperfections, and backgrounds of the scientists investigating problems at any given time.” (Buchanan et al. 2006)
- 36. Paradigms • Science is done by humans
- 37. Introduction outline – Definition of epidemiology – Epidemiologic terms – Course in context • Questions and answers • Health paradigms • Specific or general? • Recurring issues – Summary
- 38. Specific or general? • Is it reasonable to expect uniform effects of exposures on health outcomes across populations and across time? • In most instances probably not – But if not, then we are stuck with the criticism that in the face of inconsistent results we simply mobilize an auxiliary hypothesis to explain the discrepancy
- 39. Specific or general? • Income inequality and health – Controversial area of research – The greater the dispersion of income within a given society, the lower the life expectancy and the higher the mortality
- 41. Specific or general? • Theories about why this would be the case – Psychosocial interpretation – perceptions of relative disadvantage – Neomaterial – structural causes of income inequality also cause worse health outcomes
- 42. Specific or general? • Large systematic review – “…little support for the idea that income inequality is a major, generalizable determinant of population health differences within or between rich countries.” (Lynch et al. 2004) – Relations with homicide more consistent – Relations among US states more consistent • US has a unique history (slavery, segregation etc) so argued to be a product of these structural causes
- 43. Specific or general? • A recent meta-analysis was conducted suggesting that there is a consistent relation between income inequality and health across countries, but it has a threshold shape (Kondo et al. 2009)
- 44. Specific or general? • “The results also support the threshold effect hypothesis, which posits the existence of a threshold of income inequality beyond which adverse impacts on health begin to emerge.”
- 47. Specific or general? • Do we conclude that there is little support for income inequality as a generalizable determinant of health? • Do we conclude that income inequality is a generalizable determinant with a threshold shape?
- 48. Introduction outline – Definition of epidemiology – Epidemiologic terms – Course in context • Questions and answers • Health paradigms • Specific or general? • Recurring issues – Summary
- 49. Recurring issues • Struggle to get from association to cause in epidemiology – Is it even reasonable to try? – When do we make the leap? – Do we make it alone? (ie, establish causality based only on epidemiologic evidence)
- 50. Recurring issues • Documentation of risk without studying mechanism in epidemiology (“black box” criticism) – Is this informative? – Can and should we be studying mechanisms?
- 51. Recurring issues • Relationship to public health – We provide the methodologic tools for research that can inform public health decisions – We can also use tools for research that may be interesting and important scientifically but largely irrelevant to public health • Relationship to the public – Lack of understanding of what we do and how to interpret it by media and public – How can we improve this?
- 53. Move to confounding? • Discordant trial and observational study findings – Random Medical News cartoon – “apparently indiscriminate identification of particular aspects of daily life as dangerous to health” – Worst cases are when associations seen repeatedly in observational research are not realized based on trials – People observed doing certain behaviors are so different from those not doing them and the complexity of these differences not accounted for adequately with statistical procedures
- 54. Move to analyzing epi data? • Embracing and backlash against new statistical developments (last 20 yrs or so) Fancy models cannot do away with limitations of the data – Most trained in epi do not get to do original data collection – generation of “technicians” being trained – However, statistical methods available shape the questions we ask and answer and that is not a good reason to ask a particular question in a particular way
- 55. Summary • Defined epidemiology • Introduced some terminology
- 56. Summary • Perspective on where the field of epidemiology sits currently and what it struggles with • Setting the stage for the ways we will present issues in this course – Example: emphasis on causality and causal frameworks and understanding the issues in moving from association to cause – Example: emphasis on meaning of the strength of association to give better perspective on differences in associations observed across studies
- 57. Cites • Buchanan AV, Weiss KM, Fullerton SM. Dissecting complex disease: the quest for the philosopher’s stone? Int J Epidemiol. 2006;35:562-71. • Davey Smith G, Ebrahim S. Epidemiology – is it time to call it a day? Int J Epidemiol. 2001;30:1-11. • Greenland S, Gago-Dominguez M, Castelao JE. The value of risk factor (“black-box”) epidemiology. Epidemiology. 2004;15:529-35. • Rose G. Sick individuals and sick populations. Int J Epidemiol. 2001;30:427- 32. • Savitz DA. In defense of black box epidemiology. Epidemiology. 1994;5: 550-2. • Scrabanek P. The emptiness of the black box. Epidemiology. 1994;5:553-5. • Taubes G, Mann CC. Epidemiology faces its limits. Science. 1995;269:16- 169. • Weed DL. Commentary: rethinking epidemiology. Int J Epidemiol. 2006;35: 583-6.