Slide 1 Epidemiology 101 Nutritional Epidemiology Methods and Interpretation Criteria Andrew Milkowski PhD Adjunct Professor University of Wisconsin Muscle Biology Laboratory amilkowski@ansci.wisc.edu RMC 2010
Slide 2 Much Diet Health Reporting is Based on Nutritional Epidemiology Research How is this research done? What are strengths and weaknesses? The most savage controversies are those about matters as to which there is no good evidence either way -Bertrand Russell
Slide 3 Conflict Disclaimer I am not conflicted about the suitability of meat and meat products as part of a balanced and healthy human diet
Slide 4 Discussion Topics 1. Epidemiology and types of epidemiologic studies and statistics 2. Bradford Hill s Causation Criteria 3. Meat Consumption and Cancer?
Slide 5 Epidemiology Defined Epi (upon) = demos (people) + ology (study of) Historical - the study of epidemics of infectious disease Modern - the study of the distribution and determinants of health and disease frequency in human populations Epidemiology looks for patterns of disease (time, place, personal characteristics)
Slide 6
Slide 7
Slide 8
Slide 9
Slide 10 Sources of Uncertainty Definitions of foods in the study e.g. what is a processed meat Human variability Measurement error i.e. quantity consumed Uncontrolled confounding unrecognized covariates Selection and recall bias
Slide 11 Common Epidemiologic Statistics Odds Ratio (OR) = odds of exposure among ill odds of exposure among well (also called relative risk (RR) or hazard ratio (HR)) If OR (or RR) > 1, then exposure is a risk factor If OR < 1, then exposure is protective of illness If OR = 1, then there is no association between the exposure and the illness.
Slide 12 Common Epidemiologic Statistics p-value = probability that equal (or more extreme) results can be observed by chance alone; generally set at the 95% level - P < 0.05 = significant at the 95% level Confidence Interval (C.I.) Tells us about precision, accuracy and depends on: Amount of variability in the data Arbitrary level of confidence (90%, 95%, 99%) Example: OR = 2.3 (1.4-3.6) If 1.0 is not included in the C.I., the risk is significant.
Slide 13 A. Bradford Hill was a British biostatistician and epidemiologist. In 1965 he wrote an often cited article* that laid out nine criteria for evaluating statistical associations. These criteria have become an informal standard to judge epidemiological research A. Bradford Hill (1897-1991) * Hill, A. B. (1965) The environment and disease: association or causation Proceedings of the Royal Society of Medicine, 58. 295-300.
Slide 14 Don t forget Hill s own advice: None of these nine viewpoints can bring indisputable evidence for or against a cause and effect hypothesis Cited in Doll, 1991. Sir Austin Bradford Hill and the progress of medical science. British Medical Journal 305, 1521-1526.
Slide 15 The Bradford Hill Criteria 1. Strength of Association 2. Temporality 3. Consistency 4. Theoretical Plausibility 5. Coherence 6. Specificity in the Causes 7. Dose-Response Relationship 8. Experimental Evidence 9. Analogy
Slide 16 Strength of Association The stronger the relationship between the independent variable (the risk factor) and the dependent variable (the disease), the less likely that the relationship is due to something else or chance. The lung cancer rate for smokers is about 10 times (RR = 10) higher than for non-smokers. Mycotoxins in grain products are an important risk factor in liver cancer (RR~6). Many consider RR <2.0 to be highly equivocal due to inherent uncertainty in nutritional epidemiological studies.
Slide 17 Temporality The exposure must precede the disease by a reasonable amount of time, i.e., a cause must precede an effect in time. A person must smoke for years (decades) before carcinogenesis and cell transformations lead to lung cancer.
Slide 18 Consistency Multiple observations of an association, with different people under different circumstances and with different measurement instruments, increase the credibility of a causal finding. Different methods (e.g., ecological, cohort & case-control studies) produced the same result for smokers around the world.
Slide 19 Theoretical Plausibility It is easier to accept an association as causal when there is a rational and theoretical basis supported by known biological mechanisms. Smoking causes damage to the respiratory system, which over time results in cancer.
Slide 20 Coherence A cause-and-effect interpretation for an association is clearest when it does not conflict with other facts and when there are no plausible competing theories. The conclusion that smoking causes lung cancer, based on epidemiologic, laboratory animal, pharmacokinetic, clinical and other biological data, showed that all available facts stuck together as a coherent whole.
Slide 21 Specificity in the Causes Showing that an outcome is best predicted by one primary factor adds credibility to a causal claim. Lung cancer is best predicted from the incidence of smoking and chemical damage of cigarette smoke on the epithelial cell linings in the lung.
Slide 22 Dose-Response Relationship There should be a direct gradient between exposure and the risk of the disease. Data showed a positive, linear relationship between the number of cigarettes smoked and the incidence of lung cancer.
Slide 23 Experimental Evidence Related research (animal, in vitro, etc.) that is based on experiments will make a causal inference more plausible. Tar painted on laboratory rabbits ears was shown to produce cancer in the ear tissue. Hence, it was clear that carcinogens were present in tobacco smoke tar.
Slide 24 Analogy Sometimes a commonly accepted phenomenon in one area can be applied to another area. (An obtuse criterion, thus considered to be a weak form of evidence) Induced smoking with laboratory rats showed a causal relationship to tumors. A chemical may be structurally similar to another compound that is a known carcinogen.
Slide 25 World Cancer Research Fund/American Institute for Cancer Research Issued November 1, 2007 Ten Dietary Recommendations Limit red meat intake Avoid processed meat Limit salty foods
Slide 26 World Cancer Research Fund/American Institute for Cancer Research Second Expert Report Errata subsequently acknowledged
Slide 27 Meat Intake and Colorectal Cancer (EPIC Study) Norat et al., 2005
Slide 28
Slide 29 Questions to Ponder: Epidemiological Associations for Meat and Cancer Weak Associations low relative risks Inconsistencies no effect studies, gender differences Biological plausibility and coherence discussions HCA s - red meat vs. poultry, low levels in processed meats PAH s absence or low levels in processed meats Nitrite and Nitrosamines endogenous nitrogen oxide cycle Heme biological variability Confounders Vitamin C, D, aspirin intake, CLA in beef Dose response inconsistent Experimental Evidence NTP study of nitrite
Slide 30 Source: J. Heller Green Bay Gazette with permission
Slide 31 Acknowledgement Thanks to Dr. Jim Coughlin of Coughlin and Associates and Dr. Dominik Alexander of Exponent Inc. for materials used in this presentation.
Slide 32 References Boyle P., Boffetta P and Autier P. 2008. Diet, nutrition and cancer: public, media and scientific confusion Annals of Oncology 19:1665-1667 Hill A. B. 1965. The environment and disease: Association or causation? Proc. R. Soc. Med. 58, 295-300. Doll R. and Peto R. 1981. The causes of cancer: quantitative estimates of avoidable risks of cancer in the United States today." J.Nat. Canc. Inst. 66, 1191-1308. Doll R. 1991. Sir Austin Bradford Hill and the progress of medical science. Br. Med. J. 305, 1521-1526. Matanoski G. M. 2001. Conflicts between two cultures: Implications for epidemiologic researchers in communicating with policy-makers. Amer. J. Epidemiol., (Suppl.) 154, S36-S42. Phillips C. V. 2004. Analytic perspective: The missed lessons of Sir Austin Bradford Hill. Epidemiol. Perspectives & Innovations 1, 3. Rothman K. J. and Greenland S. 1998. Causation and causal inference. In Modern Epidemiology. Ed. Rothman KJ and Greenland S. Philadelphia: Lippincott-Raven, 7-28. Navia, J. L. et. al. 2010. Integrating the totality of food and nutrition evidence for public health decision making and communication Crit. Rev Food Sci Nutr. in press. National Cattlemen s Beef Association. 2010. Red meat and processed meat consumption and cancer. Accessible at www.beefresearch.org/cmdocs/beefresearch/nutrition%20research/ncba%20cancer %20Report%20100r%20Web.pdf