Nutritional epidemiology (CHS 269) Lecture NO (2) Studies in nutritional epidemiology

1. Nutritional epidemiology (CHS 269) Lecture NO (2) Studies in nutritional epidemiology • I) Observational • 2) Interventional (Experimental)

2. (1) Observational • A) Descriptive Epidemiologic studies: • 1- Case Report • 2- Case series report • 3-Correlation study (Ecological study) • 4- Cross sectional study (prevalence studies)

3. Observational (cont) • B) Analytical Epidemiological studies • Case control Study • Cohort Studies • i) Prospective cohort • ii) Retrospective Cohort study (2) Interventional (Experimental)Randomized controlled trials

4. Ecologic (correlational) Studies Compare rates of exposures and diseases in different populations using aggregate data not individual data. Usually conducted in the early stages of the investigation of a diet-disease relationship. Used for a preliminary evaluation of a new hypothesis to determine whether more extensive and costly investigations are warranted. • Because they are relatively quick and inexpensive

5. Ecologic (correlational) Studies Examples: • Very strong positive correlations were found between meat and fat consumption and the rates of breast and colon cancer in various countries. • Example, by the late 1930s, it was recognized that the use of water supplies with high fluoride concentrations led to mottling of tooth enamel. • Dentists observed that people with mottled teeth had low rates of dental caries and they hypothesized that fluoride might prevent caries. • The US Public Health Service tested this hypothesis by conducting an ecological study.

6. Researchers surveyed the dental health of children in 13 cities where the fluoride concentration in the water supply varied considerably. • The results, reported in 1942, indicated that dental caries decreased with increasing fluoride content of water. • A fluoride level of 1 ppm appeared to be optimal; • associated with a reduction in caries without an unsightly degree of mottling. • These ecological findings set the stage for later experimental studies that established the benefit of adjusting the fluoride content of drinking water supplies to one part per million.

7. Ecologic (correlational) Studies Advantages of ecological studies may be the only way to test a hypothesis • Data on a particular exposure or outcome may be available only for communities, not for individuals. • For example, it is generally not feasible to measure individual exposure to nutrients or toxic substances in drinking water. The findings for groups of people may not be applicable to individuals. 3. There are many potential determinants of disease other than the dietary factor i.e. • genetic predisposition, • environmental factors • lifestyle practices

8. Case-control studies

9. Measurement of association between exposure and outcome Odds ratios Example: Calculate the Odds ratio from the table: • Odds ratio= (60X80) ÷ (20X40)= 6 • So, high salt intake individuals are six times at risk to have hypertension than non high salt intake individuals. odds ratios greater than one indicate an increase in risk and odds ratios less than one indicate a decrease in risk.

10. Case-control studies Advantages of case-control studies. They can investigate a wide variety of potential risk factors simultaneously, Possibility to match subjects for factors that are not currently under investigation. • For example, • in a case-control study of lung cancer (profoundly influenced by cigarette smoking) • could select controls with histories of smoking as similar as possible to those of the cases • so that attention could be focused on other factors such as diet.

11. Case-control studies Disadvantages of case-control studies. Require collecting information about the subjects' past exposures They are subject to several types of bias. Various types of bias • Selection bias • Information bias • Regression dilution bias

12. Case-control studies Selection bias Selecting cases and controls from different populations Selecting control from general population by dialing random chosen phone numbers , SO, controls will not include people who do not have telephones. Because people who lack telephones tend to be of low socioeconomic status, and because socioeconomic status influences the risk of disease, the absence of these people from the control group may bias the study's findings. 2. Regression dilution bias when variability in a biological measurement leads to underestimation of the association • For example, blood cholesterol levels vary from day to day within a single individual. SO, If cholesterol levels are measured on a single occasion and then assessing the association between these levels and the risk of heart disease, the association will be weaker than the true one

13. Cohort studies

14. Relative risk (RR) From the table we can find that: • incidence of hypertension among exposed (high salt intake individuals) (Ie) a/(a+b) x 100= 20/200 x100= 10% • incidence of hypertension among non exposed (non high salt intake individuals) (I0) c/(c+d) x 100= 2/200 x100= 1%

15. Relative risk (RR) From the table we can find that: • RR= Incidence among exposed (Ie)/ Incidence among non exposed (I0) Interpretation of relative risk (RR): • If RR >1, so people exposed to a factor have an increased risk of a disease. • If RR <1, so, people exposed have a decreased risk of the disease. • If RR= 1, so there is no relation between the factor and the disease. • From this example; RR= 10/1 • SO, high salt intake individuals are ten times at greater risk of developing hypertension than non high salt intake individuals.

16. Cross Sectional (Prevalence Studies) Individuals are observed at a single point in time to provide a snapshot of the health status of the population. Some cross-sectional studies are surveys. Provide information about disease prevalence and factors associated with this prevalence. Unlike ecologic studies, it is known whether the individuals with the disease are those with the exposure. • An example, a study linked calcium intake with blood pressure measurements in healthy populations. • Data collected indicated individuals with higher blood pressure were those with lower intakes of calcium.

17. Cross Sectional (Prevalence Studies) Advantages of cross-sectional studies: • People are contacted only once, so these studies are relatively inexpensive and can be completed relatively quickly. Disadvantages of Cross-sectional studies: • Reveal nothing about temporal sequence of exposure and disease. • not known whether dietary exposure is a consequence of the disease or a causal factor. • Only measure disease prevalence, not incidence.

18. Experimental Studies (Intervention trials) Subjects are randomly assigned to receive or not receive a treatment. Ideally, the participants in a trial should not know their group assignment; through: • giving an identical-appearing but inactive placebo to those who do not receive the active treatment After a period of time, those who received the treatment are compared with those who did not, to see if their health outcomes were different. Advantages of intervention trials Providing direct evidence of a cause-and-effect relationship.

19. Experimental Studies (Intervention trials) Disadvantages of intervention trials Subjects cannot be blinded to some types of nutritional interventions. Involvement of ethical considerations that do not apply to observational studies. 3.Dietary factors may exert their effects over a prolonged period of time: For example, in a trial designed to evaluate the effect of ß-carotene supplementation on lung cancer risk in male smokers, • subjects received ß-carotene or a placebo for an average of 6 years. • Unexpectedly, who received ß-carotene showed a small increase in lung cancer risk 4. Assess only one or two factors at a time 5. Compliance with the treatment diet decreases during an extended trial

20. Confounding factors distort an association because they are associated with an exposure or a disease. unequally distributed in the study groups and may give rise to false associations. Can be avoided by analyzing data separately for subjects who fall into different categories (strata). • For example, epidemiologists analyze data separately for: • men and women, • smokers and nonsmokers • different age groups.

21. Causality In the absence of intervention trials, the following criteria are used to evaluate whether an association between an exposure and an outcome is causal. • Strength of association • Consistency of association • Dose-response relationship • Temporal relation • Plausibility • Reversibility • Study design

22. Causality Strength of association: • The stronger the association between an exposure and an outcome, the greater the likelihood that it is causal. Consistency of association • i.e. similar results have been shown in other studies. • If the association is consistently observed in different populations and under different circumstances • suggests that an exposure (factor) may be causal.

23. Causality Dose-response relationship: • If the intensity of a response increases with dose, the relation may be causal. Temporal relation: • does the exposure (factor) precede the effect (disease) • If the outcome occurs after the exposure, the relationship is truly causal

24. Causality Plausibility: • i.e. is the association consistent with other knowledge. • there is a known or hypothesized biological mechanism by which the exposure is likely to alter the risk of a disease. Reversibility: • does the removal of a possible cause lead to reduction of the disease risk? Study design: • The stronger the study design the evidence based on, the greater the likelihood the relation is causal

25. Let's carry out an epidemiological study • After attending a family dinner offering a meal composed of grilled chicken, rice, mashed potato and ice cream, 9 family members of attendance developed gastroenteritis. All attainders were interviewed. Use the data in the summary case histories to determine the food item most likely to have caused this outbreak

27. Answer