Nutritional Epidemiology Questions

1. Complete the following paragraph which describes the range of research designs that are used in nutritional epidemiology.

1. Case-control Cohort Prospective cohort Randomised Retrospective cohort Ecological 2. retrospective cohort cohort randomised case-control prospective cohort meta-analysis ecological 3. ecological retrospective cohort prospective cohort randomised cohort 4. Cohort Randomised Meta-analysis Prospective cohort Retrospective cohort Ecological 5. cohorts prospective cohorts retrospective cohorts meta-analyses ecological studies randomised-controlled studies 6. ecological analysis case-control analysis Bayesian analysis meta-analysis 7. Randomised-controlled trials Meta-analyses Prospective cohorts Retrospective cohorts Ecological studies Cohorts Nutritional epidemiologists can take a variety of approaches to examine the relationships between dietary factors and disease. 1. studies aim to recruit individuals with existing disease to compare diet and lifestyle factors with unaffected subjects from a similar population. The statistical modelling of population-level information about disease outcomes and nutritional exposures enables 2. studies to examine factors that may promote disease. In contrast 3. studies do this by collecting data at the individual level. 4. studies do this by assessing exposure levels in current populations and then following them longitudinally, whilst 5. attempt to match contemporary measurements of disease outcomes with exposure data collected at an earlier date for a different reason. In addition to making use of observational approaches, epidemiologists can employ more robust methods to examine the relationships between diet and disease. A 6. integrates evidence from a range of studies to generate more statistical power. 7. use specific interventions to examine the response of disease and disease risk factors to defined dietary exposures or food components. Types of study design 2. Note: There is negative marking for incorrect answers on this question To assess the relationship between alcohol consumption and infertility, researchers at the University of Wine recruited 42 women attending a private fertility clinic. The women were aged 35-40 and all had a history of more than 3 years unprotected intercourse without conception. For comparison a separate group of women (aged 34-41, n=50) was recruited from a local mother and toddler group. All of these women had at least one child aged between 4 and 28 months. Alcohol intake of all women was determined by use of a 5-day food diary and compared between the two selected groups. Identify the epidemiological approach taken in this study Retrospective cohort study Prospective cohort study Ecological study Randomised-controlled study Case-control study Types of study design 3. Note: There is negative marking for incorrect answers on this question The Saltiblood study was an observational study that showed an association between dietary salt, measured by urinary sodium excretion, and blood pressure. The study was based on a sample of 15000 men and women age 30-69 sampled from 75 populations spread across the world. Blood pressure data was collected by determining age-corrected average blood pressure from medical registries. Urinary sodium excretion data was collected by random sampling from each populations in the survey. Identify the epidemiological approach used in this study. Case-control study Prospective cohort study Randomised-controlled study Retrospective cohort study Ecological study Types of study design 4. Note: There is negative marking for incorrect answers on this question Researchers at the University of Littlehampton are interested in the association between adiposity and blood pressure. They recruited 75 final year engineering students (aged 20-35) for assessment of body composition and blood pressure during a single visit to their clinical research unit. Body mass index and percentage body fat were determined using a Bodpod. Blood pressure was measured using a Dynamap device. Simple regression analysis was used to explore the relationships between the measured variables. Identify the epidemiological approach used in this study Cohort study Ecological study Case-control study Randomised-controlled trial Cross-sectional study Types of study design 5. Note: There is negative marking for incorrect answers on this question To assess the putative relationship between consumption of processed red meat and colon cancer, researchers at the University of Nottingham recruited 85 patients with a confirmed diagnosis of colorectal adenocarcinoma (at least 2 years post-diagnosis) from the oncology clinics of the University of Nottingham Hospitals (all male, aged 50-76) and 85 healthy age-matched male volunteers from among the staff of the hospitals. Intake of processed red meat was determined using a recall method. Identify the epidemiological approach taken in this study Retrospective cohort study Case-control study Ecological study Randomised-controlled study Prospective cohort study Types of study design 6. Note: There is negative marking for incorrect answers on this question The relationship between antioxidant supplement use and death from cancer (all types) was assessed through follow up of 1345 men and women who had originally been recruited to the SUPPIHEART study. SUPPIHEART was designed to examine the effect of antioxidants upon blood pressure and collected data on supplement use and antioxidant supplementation using food frequency questionnaires administered in 1990, 1995 and 2000. The current study accessed medical records for the SUPPIHEART subjects up to 2012 and recorded cause of death for the 546 subjects who had died between 1990 and 2012. Identify the epidemiological approach taken in this study Retrospective cohort study Randomised-controlled study Ecological study Case-control study Prospective cohort study Types of study design 7. Note: There is negative marking for incorrect answers on this question Researchers recruited 14572 men aged 19-32 in 2001 in order to consider risk of developing hypertension in relation to intake of saturated fat. Fat intake was determined from validated food frequency questionnaires in 2001, 2005 and 2009. Blood pressure was determined at the same time as dietary assessment was completed, along with anthropometric measurements. Full data across all three timepoints was collected for 14502 men. Identify the epidemiological approach taken in this study Case-control study Retrospective cohort study Ecological study Prospective cohort study Randomised-controlled trial Types of study design 8. Note: There is negative marking for incorrect answers on this question Researchers recruited 12512 women aged 45-60 in order to consider risk of developing osteoporosis in relation to intake of calcium. All women had completed food frequency questionnaires in 1998 as part of another study (original population 18954 women, aged 30-45). The follow up (66% of original population) involved measurement of bone density using dual X-ray absorptiometry. Identify the epidemiological approach taken in this study Prospective cohort study Retrospective cohort study Cross-sectional study Case-control study Ecological study Assessment of nutritional status 9. Note: There is negative marking on this question The EURAMIC study, which investigated the link between antioxidant vitamin intake and risk of cardiovascular disease required subjects to undergo a biopsy to collect fat samples. Which of the following best describes the reason for carrying out this invasive procedure? Measurements from adipose tissue are the best measure of recent intake of fat-soluble vitamins. Fat-soluble vitamin concentrations cannot be measured in blood samples. Using a biomarker is more accurate than relying on dietary assessment tools. Measurements from adipose tissue are the best measure of longer-term intake of fat-soluble vitamins. Assessment of nutritional status 10. Note: There is no negative marking for this question Iron deficiency is one of the primary causes of anaemia. Adolescent girls are at high risk of anaemia due to irregular and chaotic dietary habits, low intakes of iron-rich foods and blood loss due to menstruation. The development of anaemia is characterised by three different stages which sequentially lead to depletion of iron stores (ferritin), reduction in availability to transport iron (transferrin receptor saturation) and insufficient capacity to maintain haemoglobin synthesis and hence red blood cell formation (full iron-deficiency anaemia). All of the options below might be suitable for assessment of iron status and anaemia. Rank them in order of appropriateness for ascertainment of habitual iron intake in a population of 3250 adolescent girls. Measurement of transferrin receptor saturation Food frequency questionnaire Measurement of haemoglobin Measurement of serum ferritin Microscopic inspection of blood smear Weighed food record 24 hour recall Assessment of nutritional status 11. Note: There is negative marking on this question 24 hour recall is a quick and inexpensive method for determining dietary intake, that is particularly useful in clinical settings. Which of the following are limitations of this method of assessing nutrient intake in populations? Tend to over-report intakes of foods consumed in low amounts Does not account for wasted food One recall is seldom representative of usual intake Is only useful when subjects are literate Tend to under-report foods consumed in high amounts Requires a high degree of cooperation from subjects Assessment of nutritional status 12. Note: There is negative marking on this question Food balance sheets are often used as the sources of dietary information in ecological studies. Balance sheets provide data on food availability and are compiled from inventories of food production, imports, exports, storage, purchases and use for non-human consumption. Which of the following statements accurately describe the strengths of food balance sheets. They reduce time, personnel and logistical constraints in surveys They are often the best data available on national food consumption practices They provide a good indication of food habits and dietary trends for a country They are practical way of estimating nutrient intakes in large populations Nutritional assessment 13. Note: There is negative marking on this question When evaluating dietary intakes researchers need to balance decisions about accuracy and practicality of methodology in deciding the ideal approach for their study. Typically, where food or nutrient intakes are required as exposure measures the decision will be made to adopt either an estimated food record, weighed food record or food frequency questionnaire. Each of these have advantages in particular circumstances. 1. the weighed food record the food frequency questionnaire the estimated 2. the estimated food record the food frequency questionnaire the weighed 3. the weighed food record the estimated food record the food frequency 4. the weighed food record the food frequency questionnaire the estimated 5. the estimated food record the weighed food record the food frequency 6. the estimated food record the food frequency questionnaire the weighed 7. the food frequency questionnaire the weighed food record the estimated 8. the estimated food record the food frequency questionnaire the weighed 9. the estimated food record the weighed food record the food frequency 10. the e food the w food the f frequ stimated record eighed record ood ency food record food record questionnaire food record questionnaire Complete the statements below: Researchers wishing to determine micronutrient intakes in a large population would prefer to use 1. Researchers wishing to determine habitual food intakes in a large populations would prefer to use 2. Researchers wishing to determine food and nutrient intakes in a small group would prefer to use 3. The most accurate approach to measuring dietary intake is 4. approach. The 5. approach is the most expensive method for assessing intake. The 6. approach is most subject to recall bias. The 7. The greatest burden upon the subjects is associated with 8. The lowest burden on the researchers is associated with 9. approach. The 10. approach is widely considered to be the method of choice for investigating diet-disease relationships. Levels of evidence 14. The pyramid diagram shows the hierarchy of epidemiological studies, with the study designs that are best able to infer causality of associations between nutrition and disease at the top, and the designs with least power to infer causality at the base. Correctly label designs A-F. food record food record food record quest approach. approach. approach. approach gives the least accurate representation of portion sizes chosen by respondents. questionnaire ionnaire Randomised-controlled trial Cohort study Case-control study Ecological study Cross-sectional study Meta-analysis i. A ii. B iii. C iv. D v. E vi. F Sources of error 15. Note: There is no negative marking on this question A survey considering the relationship between obesity and physical activity included centres in eight different regions of Britain. A stronger relationship was noticed in the North of England and Scotland, which might suggest a greater degree of risk in these regions. Closer investigation, however, indicated that the prevalence of obesity was similar in all regions and that it was the interviewers approach to recording physical activity data that varied between regions. Inferring relationships between obesity and activity levels from this survey would therefore be inappropriate. This provides an example of what inherent defect of some epidemiological investigations? Sources of error 16. Note: There is no negative marking on this question. Researchers had suggested that individuals who have a high intake of cruciferous vegetables have a lower risk of osteoarthritis than those who have lower intakes of these foods. Their conclusion was reached by measuring the vegetable intakes of 150 randomly selected men and women (aged 40-75) in each of 32 NHS hospital trust areas and using regression analysis of this data against average prevalence data for osteoarthritis in each of those areas. However, several retrospective cohort studies have since demonstrated no association between cruciferous vegetable intake and risk of developing osteoarthritis. The conflict in the evidence base detailed in the above scenario is likely to reflect a particular problem associated with the design of the first epidemiological study described. What is the term for this particular defect? Sources of error 17. Note: There is negative marking on this question Researchers at the University of Inverhampton investigated the association between consumption of green tea and risk of oesophageal cancer by recruiting 143 patients attending the oncology clinic of Inverhampton General Hospital for treatment and 140 patients attending the same hospital for other conditions. Recruitment: Cancer patients were selected on the basis of a firm medical diagnosis (based upon CT scan) of either adenocarcinoma or squamous cell cancer of the oesophagus. The patients were aged between 40 and 65 and comprised 52 women and 91 men. Stage of disease was variable with 31% newly diagnosed, 49% in treatment for at least 6 months and 20% undergoing palliative care. Controls were all selected from other areas of the hospital and were matched to the cancer patients for age (within 6 months), sex, weight, postcode and ethnicity. Controls had been admitted for a variety of conditions including peptic ulcers, coeliac disease, fractures, dental surgery, cardiovascular and renal disease. Subjects were excluded from taking part as a control if they had ever had a cancer diagnosis. Assessment of tea intake: Intake of beverages was determined using a modification of a questionnaire that was previously validated (Johnson and Hicks, 2004, Br J Beverages 11, 666-667) for determination of commonly consumed beverages in the UK adult population. Questions included a range of teas (black tea, green tea, oolong tea, herbal tea) and intake was quantified on the basis of typical cups per day (a cup=250ml) consumed in the last two years. Confounding factors: Data was analysed by binary logistic regression modelling to calculate odds ratios for oesophageal cancer, adjusted for confounding factors in the study. Variables considered as confounding were sex, age, socioeconomic status, BMI, height, menopausal status (women), abdominal circumference, medication, cancer treatment, hydration status, antioxidant capacity, surgical history, smoking history and family history of cancer. Having reviewed the data obtained in this study the researchers concluded that there was no evidence to support the claim that consumption of green tea prevented oesophageal cancer. The design of the study raises a number of questions that might challenge the validity of the conclusion. Which of the following statements about sources of error are true? Selection Matching of the patients and controls according to key characteristics was thorough and increased the likelhood of observing a causal relationship between exposure and outcome. 1. Matching the subjects on the basis of ethnicity is appropriate as it is a confounding factor. 2. Matching the subjects on the basis of weight is appropriate as it is associated with risk of oesophageal cancer. 3. Matching the patients and controls on several factors made the study more efficient as complex adjustment in analysis was unnecessary. 4. Sufficient care has been taken to ensure that controls were not suffering from oesophageal cancer. 5. The control population is innappropriate as the subjects are unhealthy. 6. Measurement Recall bias has been minimised by limiting assessment of beverage intake to the last two years. 7. The validation of the beverage questionnaire means that it is appropriate for estimating intake of green tea. 8. Ascertainment of the cancer cases was robust, providing a good basis for the investigation of risk factors. 9. Inclusion of the two main forms of oesophageal cancer is a strength of the study as it increases the power to uncover an association with the exposure. 10. 1. Partly True True False 2. True False Partly True 3. True False Partly True 4. True False Partly True 5. True False Partly True 6. True Partly True False 7. True False Partly True 8. True Partly True False 9. True False Partly True 10. True Partly True False 11. False True Partly True 12. False Partly True True 13. Partly True False True 14. False True Partl True y Confounding By matching the patients and the controls, the study has made adjustment for confounding factors unnecessary. 11. The range of confounding variables that were included in the analysis is comprehensive and there are no concerns about confounding in the study. 12. Adjustment for for sex and age is unnecessary as these variables were used to match the patients and controls during selection. 13. Adjusting the data for BMI was necessary as matching by weight did not make allowance for body composition. 14. Over-adjustment for confounders in this study represents a type of bias as it can mask a relationship between exposure and outcome that is really present. Sources of error 18. Note: There is negative marking on this question The research team of Professor Maddy McMad of the University of Western Kyrgyzstan have published the findings of their study into the relationship between trans-fatty acid intake and myocardial infaraction (YAKMIC) in the Journal of Nutritional Sensations. The YAKMIC study was a prospective cohort study which originally included 22345 adult men (all ethnic Uzbeks, aged 32-45, men age 34.7yrs) who were recruited in 2001. At recruitment the men had a thorough clinical examination to ascertain cardiovascular health (blood pressure, electrocardiogram, carotid-intima thickness measurement, blood lipids including circulating trans-fatty acids), metabolic indicators (insulin, glucose, HOMA-IR) and anthropometric indices (BMI, abdominal circumference, body fat by triceps and sub-scapular skinfolds) and current intakes of dietary fats were estimated using a food frequency questionnaire (FFQ) validated for the local population (McMad et al, 2000, Journal of Obscure Research, 43 555-582; showed that agreement between FFQ and weighed food record was high for saturated fats, trans-fatty acids, PUFA and MUFA R2=0.89, 0.91, 0.82, 0.75 respectively). Anthropometric measurements and FFQ records were administered by the team of 22 researchers comprising the YAKMIC group (variation in measures baseline clinic: height 5.2%, weight 4.3%, triceps skinfold 11.2%, sub-scapular skinfold 12.5%; final clinic: height 1.1%, weight 2.2%, triceps skinfold 5.2%, sub-scapular skinfold 4.3%). Comparison of FFQ data and measured blood lipids showed a high degree of agreement (R2=0.92)The typical diet of this population comprised meat and dairy produce, with limited intakes of processed foods. Saturated fat intakes were typically high (on average 15% of daily energy intake). Trans-fatty acid intakes were primarily from dairy sources. Participants were followed up in 2003, 2005, 2007 and 2009 and at each point all of the baseline measures of cardiovascular health and fat intake were repeated. At the final follow-up clinics in 2012 the full medical history of the men over the 11 years of the study was accessed and cardiovascular events (myocardial infarction, angina, arrythmia) recorded. YAKMIC was based in a largely rural area giving a stable population that was easily followed across the duration of the study. Follow-up to 2009 was 92.3%. However, local conflicts during the period 2009-2011, led to the deaths of a significant number of subjects in military service and displacement of populations in some regions. As a result participants were unavailable to attend all of the clinical follow-ups. As a result the analysis was confined only to the population that had been able to attend all 6 assessments (3075 men, 13.75% of original cohort). Analysis of the baseline characteristics (BMI, adiposity, cardiovascular and metabolic measures) of the subjects who completed the full study showed no major systematic difference between completers and subjects lost to follow-up (except that non-completers were significantly younger: completers average age at recruitment 37.5, non-completers 32.2 years). Among the completers there were 462 cardiovascular events reported. The analysis of the data collected in this study suggested that there was a strong inverse association between intake of trans-fatty acids and myocardial infarction in this population. Critics of the study have raised a number of concerns about the study design as listed below. Indicate which of these criticisms you consider to be valid and detract from the researchers interpretation of the findings. The population is not representative of western populations so findings cannot be widely generalised. 1. The study was of insufficient duration to observe a relationship between fat intake and cardiovascular disease. 2. Insufficient data was available to adjust for important confounding factors. 3. The high drop out from the study (87%) makes the findings unreliable.4. Estimation of fat intake on the basis of a food frequency questionnaire introduces unacceptably high measurement error. 5. Selection of just the participants with a complete set of data introduces bias as the relationship between fat intake and disease may be different when considering partial data sets. 6. 1. Partly valid Not valid Valid 2. Valid Partly valid Not Valid 3. Not valid Valid Partly valid 4. Not valid Valid Partly valid 5. Valid Not valid Partly valid 6. Partly valid Valid Not valid 7. Valid Partly valid Not valid 8. Partly valid Valid Not valid 9. Partly valid Valid Not valid 10. Not valid Partly valid Valid There was a marked variability in the measurement of anthropometric indices between researchers, introducing significant information bias to the study. 7. The conclusions of the study are the opposite of most earlier reports (positive association between trans-fats and disease) so a causal relationship should not be inferred. 8. The conclusions of the study are the opposite of most earlier reports (positive association between trans-fats and disease) the YAKMIC study is of no value to researchers in this field. 9. In a diet with little processed food, the population would not be appreciable quantities of trans-fatty acids, limiting the power of the study to examine an association. 10. Confounding and bias 19. Note: There is negative marking for incorrect answers on this question Which of the diagrams most accurately represents the confounding effect of smoking upon the relationship between alcohol consumption and heart disease. A B C D Confounding and bias 20. Note: There is negative marking for incorrect answers on this question The association between insulin resistance and body composition was determined in a group of 145 men aged 55 and 64. Insulin resistance was determined using the HOMA-IR score (derived from circulating insulin and glucose concentrations) and body composition was measured by bioimpedance. A range of potential confounding factors was also considered. The table below shows the results of simple analysis to determine the factors that were associated with insulin resistance and body fat mass. Table: Regression coefficients showing the relationship between HOMA-IR score and other variables measured in the study. Variable Relationship with HOMA-IR Relationship with % body fat Socioeconomic status -0.120* -0.156* Age (yrs) 0.060 0.216* Sedentary occupation 0.110* 0.321** Vigorous physical -0.098 -0.111 activity (hrs/wk) Alcohol consumption 0.085 0.142 (units/wk) Cigarette smoking 0.021 0.148* (cigs/d) Mean energy intake 0.197** 0.251** (MJ/d) Energy from fat (%) 0.221** 0.156* Energy from 0.084 0.202** carbohydrate (%) Energy from protein 0.023 0.101 (%) Data are shown as r values determined by Pearson’s regression analysis. * denotes P<0.05, ** P<0.01 Considering the scenario presented, which variables should be adjusted for as confounding factors in the data analysis? Energy intake, physical activity, % energy from fat, cigarette smoking Social class, sedentary occupation, energy intake, % energy from fat Social class, sedentary occupation, energy intake, % energy from carbohydrate Energy intake, sedentary occupation, age, alcohol consumption Confounding and bias 21. Note: There is negative marking for incorrect answers on this question An investigation into the association between green leafy vegetable consumption and risk of breast cancer considered components of dietary intake and disease in a prospective cohort study involving 4300 women. In the analysis of the data the researchers considered the role of potential confounding factors as set out in the diagram below. Regression analysis was then used to assess the relationship between each of the identified factors and the primary exposure and outcome variables in the study. These analyses are shown in the table below. Variable Relationship with Relationship with primary exposure primary outcome Age (yrs) 0.172* 0.253** Menopausal status 0.104* 0.152* Socioeconomic class 0.165 0.102* Number of children 0.025 0.265** Meat intake (g/d) -0.201* 0.365** Folic acid intake 0.452** -0.325** (ug/d) Data are shown as r values determined by Pearson’s regression analysis. * denotes P<0.05, ** P<0.01 Using the information presented, identify the factors that should be adjusted for as confounders in the analysis of the data. Age Folic acid intake Meat intake Menopausal status Number of children Socioeconomic status Confounding and bias 22. Note: There is negative marking on this question The extract below is taken from a published study that is an element of the EPIC study, a prospective cohort study investigating the nutritional risk factors for cancer. The EPIC (European Prospective Investigation into Cancer and Nutrition) is a multicenter, prospective cohort study that investigated the role of metabolic, dietary, lifestyle, and environmental factors in the development of cancer and other chronic diseases. Briefly, between 1992 and 2000, 521,448 volunteers aged between 25 and 70 years were recruited in 23 centers from 10 European countries (Denmark, France, Germany, Greece, Italy, Netherlands, Norway, Spain, Sweden, and the United Kingdom). In France, Norway, Utrecht (Netherlands), and Naples (Italy), only women were included. In general, individuals were selected from the general population of a specific geographic area, a town or a province. Exceptions included the French cohort, based on members of the health insurance system for state-school employees, and the Utrecht cohort, based on women who underwent breast cancer screening. Participants were invited to participate either by mail or in person and those who agreed to participate signed an informed consent agreement. Approval for this study was obtained from the ethical review boards of the International Agency for Research on Cancer and from all local institutions. After the exclusion of individuals without dietary and nondietary questionnaires, individuals without data on weight and height at baseline or with extreme or implausible anthropometric values, pregnant women, and those in the top and bottom 1% of the ratio between energy intake to estimated energy requirement, 497,735 individuals were available for the analyses. We further excluded 123,932 subjects without weight data at follow-up or with extreme or implausible weight changes. Thus, 373,803 subjects (103,455 men and 270,348 women) were included in the present analysis. Assessment of anthropometric measures and weight change Two weight measures were available for each participant: one measure at baseline and one at follow-up. In most centers, body weight and height were measured at baseline with the use of similar, standardized procedures. The exceptions were the centers of Oxford (United Kingdom), France, and Norway, where self-reported anthropometric values at baseline were collected. Self-reported weight was obtained at follow-up in all centers, except in Norfolk (United Kingdom) and Doetinchem (Netherlands), where weight was measured. The accuracy of self-reported anthropometric measures was improved with the use of prediction equations derived from subjects with both measured and self-reported measures. Body mass index (BMI; in kg/m2) at baseline was calculated as weight in kilograms divided by height in meters squared. Because the follow-up times differed by center (from 2 y for Heidelberg to 11 y for Varese), our main outcome is the annual weight change (g/y; ie, weight at follow-up minus weight at baseline divided by time of follow-up (in y). Because the association of weight change with meat intake was similar with the use of self-reported or predicted weight, all the results presented here are based on predicted weight at follow-up. Dietary assessment Usual dietary intake at baseline was measured with the use of country-specific validated questionnaires. Most centers adopted a self-administered quantitative dietary questionnaire of 88–266 food items. Semiquantitative food-frequency questionnaires were used in Denmark, Norway, Naples, and Umeå (Sweden). Combined dietary methods were used in the United Kingdom and Malmö. Nutrient intakes were calculated with the use of the EPIC Nutrient DataBase, a standardized food-composition table. To adjust for possible systematic under- or overestimation in dietary intake measure, a dietary calibration study was conducted with the use of a random sample of ≈36,900 men and women who completed an additional computerized 24-h dietary recall (EPIC-SOFT). In this study, total meat includes red meat (beef, veal, pork, and lamb), processed meat (ham, bacon, sausages, and other meat products mainly from beef and pork), and poultry (mainly chicken and in some cohorts, turkey and rabbit). Assessment of other covariates Lifestyle and health factors were collected by questionnaires at baseline. These included questions on tobacco smoking, educational attainment, physical activity (occupational and leisure-time activity), menstrual history, and history of previous illness. Smoking tobacco was also collected during follow-up, at the same time as the anthropometric measures. This permitted smoking status modification to be taken into account during follow-up in our analysis. 1. True False 2. False True 3. True False 4. True False 5. False True 6. False True 7. True False 8. False True 9. True False 10. discriminatory power external validity statistical significance internal validity 11. confounding bias information bias recall bias measurement error sampling bias 12. reliance of self-report measurements recruitment of only women from some centres variation in measurement methods between centres failure to collect biomarkers of intake exclusion of people with missing data Having considered the extract, answer the following True/False questions (1 mark each). The use of food frequency questionnaires is widely regarded as a gold-standard approach for measuring nutrient intake in large populations. 1. Self-report of tobacco smoking behaviour is a potential source of social desirability bias. 2. Information bias is only of significance to a study if it results in misclassification of the exposure or outcome. 3. Information bias and measurement error are distinct sources of bias in cohort studies. 4. Monitoring changes in self-reported smoking behaviour increases the risk of introducing bias into the analysis. 5. Social desirability bias is a component of information bias that is a major issue with self-reported data. 6. Self-reported anthropometric measurements are a source of information bias, but this bias will be random across the population. 7. The loss of 28% of the cohort represents a major confounding factor in this study. 8. The adjustment of the weight change data for the time of follow-up removes any information bias associated with the centres not having the same follow-up periods. 9. Complete the following paragraph, selecting from the options listed (1 mark each option). The EPIC study has strong 10. by virtue of the large population recruited across a number of different centres. There were inconsistencies in recruitment between the centres however, which has introduced some 11. . Most 13. outcome exposure important among these was the 12. . The study recorded intake of meat as a primary 13. variable and made use of 14. to measure this variable. The 15. variable was weight gain, which was measured by self-report. The use of prediction equations to adjust self-reported weights was a feature of the data analysis and this reduced the . 16. Confounding and bias 23. Note: There is no negative marking for this question The extract below is from a paper considering the relationship between ApoE genotypes and risk of prostate cancer. Method and data collection This report examines the association between the ApoE4 allele by ethnic or country background, dietary supply factors, and per capita gross domestic product (GDP) for males with prostate cancer incidence and mortality rates for 122 countries. Values of ApoE4 by country or ethnic background were obtained from data for individual countries, as well as several summaries of such values. For Europe, the data were graphed with respect to the latitude of the population center for each country, and the regression value was used for each country. For countries with mixed ethnicity, the value was a proportional combination of the values for the ethnic backgrounds in the countries of origin. The uncertainties in the ApoE4 prevalence values are estimated to be 10%-25% of the value. One contribution to the uncertainty is that ApoE4 prevalence is not a well-determined value in that it is determined from targeted studies. For several countries, no determinations were made, so estimates were made on the basis of values from other countries. It was noted that for ApoE4 prevalence below 30%, prostate cancer rates generally increased but that rates were very low for countries with prevalence of 30% or higher. This effect is similar to that observed for the prevalence of Alzheimer’s disease and largely relates to dietary differences between developing and Western developed countries. Thus, a subset of 102 countries with ApoE4 prevalence less than 30% was used in a second regression analysis. Per capita GDP has been suggested as a risk-modifying factor for several types of cancer. Data for per capita GDP for 2000 were obtained from an almanac. Because of the large range of GDP, the square root of the value was used in the regression analyses to reduce the effect of extreme values. The prostate cancer data used are incidence and mortality rates for 2002 from GLOBOCAN 2002. The criteria for inclusion were that the population of the country had to exceed 250,000 inhabitants, that the data had to be of high quality, that dietary supply data had to be available, and, generally, per capita GDP data had to be available for 2000. Data for several other types of cancer were also obtained to investigate the findings for various risk-modifying factors for prostate cancer: breast, colorectal, endometrial, ovarian, pancreatic, and renal cancer, with lung cancer incidence data for 2002 used to determine the role of smoking in risk for each type of cancer. Dietary supply data were obtained from the Food Balance Sheets of the FAO. These data represent food disappearance in the population. Although food consumption by individuals accounts for only about 70% of the food that disappears, the factor is similar in most populations and therefore serves as a reliable index. The data examined in this study were animal and vegetable fat, cereals, fish, milk, milk protein, milk fat, onions, sweeteners (added sugar), and tomatoes. Only animal fat, cereals, and milk protein were studied in detail because the other factors did not have significance in preliminary analyses for prostate cancer. 1. cohort cross sectional ecological 2. high low moderate 3. they only show intake among adults in the population they record overall availabilty of commodities per head of population rather than intake they do not discriminate between consumption and storage they use data that is collected at different times in different populations Complete the blanks below: This is an example of a 1. study and as such has 2. capacity to determine a causal relationship between genotype, dietary factors and risk of prostate cancer. The data on intake of foodstuffs was estimated using the FAO food balance sheets. These data are prone to error because 3. . The external validity of the study is enhanced by 4. . The estimates of ApoE4 prevalence in countries where no genotyping had been performed could introduce 5. . 4. the inclusion of 122 different countries with wide variation in risk the use of robust GLOBOCAN cancer data the use of robust FAO data the inclusion of developed and developing countries 5. selection bias information bias confounding bias Definition of key terms 24. Note: There is no negative marking for incorrect answers on this question An ecological fallacy is a misinterpretation of epidemiological data where researchers attribute the nature of individual-level data to larger populations. True False Definition of key terms 25. Note: There is negative marking for incorrect answers on this question Which of the following is the correct definition of a Type I error in statistical analysis. Incorrect acceptance of the null hypothesis when it is false. A false negative result. Correct acceptance of the hypothesis, but with the direction of effect wrongly attributed. Incorrect rejection of the null hypothesis when it is true. Definition of key terms 26. Note: There is no negative marking for incorrect answers on this question Neyman’s bias is also referred to as Prevalence-incidence bias. This is a form of selection bias in case-control or cross-sectional studies that occurs due to selective survival among individuals who should be included as cases. This occurs because they have recovered or died. True False Definition of key terms 27. Note: There is negative marking for incorrect answers on this question Complete the paragraphs below. 1. correlation coefficients standard errors confidence intervals standard deviations 2. common diseases in all cases rare diseases 3. Relative Risk Odds Ratios 4. Relative Risk Odds Ratio 5. the numbers of people in each of two groups that experience an event the probability of experiencing an event in each of two groups 6. show the risk of developing a disease following a specific exposure compare the odds of a disease occuring in one group relative to another 7. the probability of experiencing an event in each of two groups the numbers of people in each of two groups that experience an event 8. shows the risk of developing a disease following a specific exposure compares the odds of a disease occuring in one group relative to another Studies in nutritional epidemiology often report outcomes using odds ratios, relative risk and hazard ratios as statistical terms. These provide key indicators of the strength of association between disease and nutritional factors and are interpreted alongside 1. . Odds ratios and relative risk are essentially similar for 2. but are used in different ways according to specific epidemiological designs. 3. are the usual outcome reported in case-control studies, whilst randomised controlled trials and cohort studies more usually report the 4. . Odds Ratios are calculated from and specifically 6. . Relative Risk is calculated from 7. and specifically 8. . 5. Definition of key terms 28. Note: There is no negative marking for incorrect answers on this question Complete the definitions below. 1. Total number of people in the population Total cases of disease 2. Total number of people in the population Age standardized rate 3. Morbidity rate in a study group Total cases of disease 4. Morbidity rate for a reference population Total number of people in a population 5. Morbidity rate in a study group Total number of people in the 6. Total number of people in the population Morbidity rate for a 7. disease incidence disease prevalence confounding effects of such as age and sex Age adjusted cases of disease for the population Total cases of disease Morbidity rate for a reference population Morbidity rate in a study group year of measurement population Morbidity rate for a reference population Total cases of disease reference population Morbidity rate in a study group Total cases of disease The Crude Morbidity Rate is calculated as the 1. / 2. . The Standardized Morbidity Ratio is calculated as the 3. / 4. . The Standardized Morbidity Rate is calculated as the 5. / 6. , adjusted for 7. . Incidence and prevalence 29. Note: There is negative marking for incorrect answers on this question The Centre for Obesity studies has monitored the weight of all children who started school in the city of Liverchester in 2003 over a ten year period. In total 15000 children were included in the monitoring programme. At the start of the study 4.3 % of the children (n=645) were classified as obese using the International Obesity Task Force criteria. By 2013 this had increased to 1650 children (11%). Which of the following statements are correct? The prevalence of obesity in 2013 was 1100 cases per 10000 children. The incidence of obesity in 2001 was 0.43 per 1000 children. The incidence of obesity over the 10 year period was 110 cases per 1000 children. The prevalence of obesity over the ten year period was 6.7%. The prevalence of obesity in 2013 was 70 cases per 10000 children. The incidence of obesity over the ten year period was 67 cases per 1000 children. Relative risk and confidence intervals 30. Note: There is no negative marking for incorrect answers on this question A population of 20420 men completed food frequency questionaires in 2008 and were followed up in 2013 to assess the incidence of lung cancer over the subsequent five year period. The relative risk of lung cancer associated with fat intake in excess of 80g/day was 2.24 (95% CI 1.45, 3.03), with fruit and vegetable intake over 400g/day was 0.53 (95%CI 0.05, 1.01) and with meat consumption in excess of 150g/day was 1.83 (95%CI 0.93, 2.73). A greater intake of fruit and vegetables is associated with a reduced risk of lung cancer in this population. True False Interpretation of data 31. A population of 256 seven-year old children from the UK was recruited from two general practitioner surgeries. Historical anthropometric data from birth to age three and current anthropometry were assessed as predictors of parent- reported wheeze (asthma) and eczema. Social class was recorded on the basis of paternal occupation (social class I= highest class, social class V= lowest class). The table below shows the analysis of the data, indicating the odds of a child developing a wheezy chest (asthma) by the age of 7 years associated with anthropometric measures and potential confounding factors. Factor Odds ratio 95% confidence interval Head circumference over 35.5cm between 0.91 0.84-0.98 10 and 15 days of age Ponderal index 1.32 1.01-1.63 >13.98 at 7 years Current passive 1.57 0.96-2.07 smoking exposure Exposure to smoking 4.12 0.87-6.95 in pregnancy Lower social class 0.45 0.02-1.32 Male sex 0.92 0.84-1.05 A family history of 7.25 0.92-15.86 wheeze Complete the matrix below to explain the relationship of each of the factors with risk of developing a wheezy chest at age 7. Reduced risk Increased risk No effect on risk i. A family history of wheeze ii. Male sex iii. Lower social class iv. Exposure to smoking in pregnancy v. Current passive smoking exposure vi. Ponderal index >13.98 at 7 years vii. Head circumference over 35.5cm between 10 and 15 days of age Interpretation of data 32. Note: There is negative marking for incorrect answers on this question In a study of the relationship between meat intake and risk of breast cancer, it was noted that the relative risk of cancer associated with a red meat intake of 150 g/d was 1.25 (95% CI 1.13-1.37) compared to consumption of 0-50 g/d. Which of the following statements is correct? Consuming red meat at more than 150 g/d significantly increases risk of breast cancer by 125% Consuming red meat at more than 150 g/d significantly decreases risk of breast cancer by 25% Consuming red meat at more than 150 g/d has no significant effect on risk of breast cancer Consuming red meat at more than 150 g/d significantly increases risk of breast cancer by 25% Interpretation of data 33. Note: There is no negative marking for incorrect answers on this question A study evaluating the relationship between stomach cancer and diet evaluated intake of foods and beverages using a food frequency questionnaire in 1200 cancer patients and 1197 controls. Several factors were found to be significantly associated with cancer risk. Food item RR for stomach 95% CI cancer Meat <100 g/d >100 g/d 1 2.52 2.42-2.62 Fish <20g/d >20g/d 1 0.69 0.49-0.89 Salt <6g/d >6g/d 1 2.56 1.56-3.56 Alcohol <20g/d >20g/d 1 3.57 3.03-4.14 Vegetables <250g/d >250g/d 1 0.52 0.32-0.72 Garlic <5g/d >5g/d 1 0.24 0.11-0.37 Considering the data in the table, rank the following factors in order of the strength of their effect upon risk of developing cancer, with the strongest pro-cancer agent first and the strongest anti-cancer agent last. Fish intake Alcohol intake Salt intake Vegetable intake Garlic intake Meat intake Interpretation of data 34. Note: There is negative marking for incorrect answers on this question A cohort of 845 children aged 6 to 8 years were recruited for an assessment of the relationship between sugar intake from sugar sweetened beverages (SSB) and overweight. All children were weighed and height measured to determine BMI. Children were then classified as normal weight, overweight or obese based upon BMI centiles. Intake of SSB was determined from a food frequency questionnaire. The study identified sex, social class and parental weight as confounding factors and adjusted the analysis accordingly. Unadjusted Model 1 Model 2 Model 3 Model 4 SSB0-50 1 1 1 1 1 g/d SSB 51- 100 g/d SSB >101 g/d 1.2 (0.9- 1.5) 1.8 (1.5- 2.1) 1.2 (0.9- 1.5) 1.9 (1.6- 2.2) 1.3 (0.9- 1.6) 1.9 (1.6- 2.2 1.8 (1.2- 2.4) 2.2 (1.6- 2.8). . 1.7 (1.3- 2.1) 2.3 (1.8- 2.8) All data are shown as OR (95% CI) for overweight and obesity. Model 1- adjusted for confounding effects of child sex Model 2- adjusted for confounding effects of social class Model 3- adjusted for confounding effects of parental weight Model 4- adjusted for all confounders Based on this data it can be stated that sugar sweetened beverages significantly 1. the risk of overweight and obesity in children when consumed at more than 2. . The analysis considered how 3. impacted on the relationship between consumption and overweight and showed that 4. was the most potent confounder. The effect of sugar sweetened beverages upon risk of overweight was greatest when 5. considered in the model and overall the risk was 6. greater in children consuming more than 100g/d than in children consuming 0-50g/ d. Interpretation of odds ratio data 35. Note: There is negative marking for incorrect answers on this question The table below shows data from a case-control study which considered the relationship between intakes of fruits and vegetables and risk of esophageal cancer. The study population included a total of 38,790 men aged 45-74 years, among which 116 cancer cases were detected. Lifestyle characteristics were investigated using a self-administered questionnaire, which included a validated food frequency questionnaire with 138 food and beverage items. The data are presented as hazard ratios with 95% confidence intervals. 1. decrease increase 2. 0g/d 50g/d 100g/d 3. confounding factors social class sex parental weight 4. parental weight social class sex 5. just parental weight was just social class just sex was all confounders were 6. 2.3-fold 1.9-fold 2.2-fold 1.8-fold Tertiles of intake Fruit &Veg combined Lowest Middle Highest P for trend Per 100g/day Median Hazard intake ratio (g/day) 170 1 322 0.85 544 0.52 0.56-1.29 0.30-0.88 95% CI 0.02 0.89 0.79-0.99 Based upon the data presented for this study, which of the following statements are true. Increasing intake of fruit and vegetables from 100 to 200g/day would reduce risk of esophageal cancer by 11%. Vegetables had no protective effect against esophageal cancer. Consuming fruit and vegetables in accordance with the 5-a-day guidance would be sufficient to attain significant reduction of risk of esophageal cancer. The beneficial effects of fruit and vegetables was mostly attributable to a protective role of fruit. Individuals in the highest tertile for fruit and vegetable intake were 48% less likely to develop esophageal cancer. Interpretation of data 36. Note: There is negative marking for incorrect answers on this question The manufacturer of nutritional supplements, Slangtech, has been trialling three new products which may have the capacity to increase fertility in women. Evansone: 40 mg/d ascorbate, 5 mg/d folic acid Fertisup: 40 mg/d ascorbate, 10 mg/d zinc, 20 mg/d iron, 150mcg/d vit B6 Ovusup: 40 mg/d ascorbate, 20mg/d carnitine, 20mg/d taurine For each trial the company recruited 100 married women who had been infertile for a period of at least 12 months, who were not receiving assisted reproductive treatement. In each trial 50 women received a daily test supplement and 50 received a placebo. All women were followed up after 6 months and the numbers of conceptions in each group were recorded. The results are shown below: Evansone trial: Odds of pregnancy for supplemented women vs placebo 1.32 (95% CI 1.12-1.44) Fertisup trial: Odds of pregnancy for supplemented women vs placebo 7.66 (95% CI 0.95-14.37) Ovusup trial: Odds of pregnancy for supplemented women vs placebo 0.84 (95% CI 0.70-0.98) Fruit only Lowest 47 1 Middle 137 0.8 Highest 280 0.65 P for trend 0.09 Per 0.90 100g/day Veg only Lowest 88 1 Middle 165 0.8 Highest 286 0.68 P for trend 0.10 Per 100g/day 0.52-1.23 0.39-1.08 0.76-1.07 0.52-1.23 0.42-1.10 0.81 0.66-0.98 1. 2. Fertisup Ovusup Evansone Evansone Ovusup Fertisup Which of the three products would you advise Slangtech to develop further towards a commercial launch? 1. . Of the three products, one was found to carry a risk of reducing conception rates among the women. Which product would you advise Slangtech to cease developing? 2. ( Calculation of odds ratio for colon cancer 37. Note: There is no negative marking for incorrect answers on this question A case control study investigating the relationship between dietary fibre and colon cancer initially recruited a population of 1262 men with a confirmed diagnosis of colon cancer and 1320 men with no evidence of colon cancer. All subjects completed a food frequency questionnaire that examined habitual food intake over the previous 5 years. Complete records were obtained from 1243 men with cancer and 1298 men without cancer. From the dietary data all individuals were classified as having a high fibre diet (>20g/day) or a low fibre diet (<20g/day). 450 men with cancer and 832 men without cancer were classified as high fibre consumers. Calculate the odds ratio for a diagnosis of colon cancer associated with the consumption of a low fibre diet in this populations. You may wish to draw a 2×2 contingency table to help with your calculations. The calculation of odds ratio is based upon the formula (axc)/(bxd), where ‘a’ and ‘c’ are the number of subjects in each cell of the table which is in agreement with the hypothesis. And ‘b’ and ‘d’ are the number subjects in each cell of the table which is in disagreement with the hypothesis. The calculation is ( * )/( * ) The odds ratio is Calculation of risk of hypertension 38. Note: There is no negative marking for incorrect answers on this question A population of 3000 men were recruited for a study of the relationship between sodium intake and blood pressure. All men had blood pressure determined using a mercury sphygmanometer and were classified as normotensive (systolic blood pressure less than 140 mm Hg) or hypertensive (systolic blood pressure 140 mm Hg or higher). Sodium intake was determined from 7-day food records and the men were classified into 5 groups based on intake. The table below shows the distribution of normotensive and hypertensive men in the 5 groups. Sodium intake (g/d) Normotensive, n (%) Hypertensive, n (%) 0-2.9 525 (17.5) 108 (3.6) 3.0-5.9 407 (13.6) 122 (4.1) 6.0-8.9 650 (21.7) 311 (10.4) 9.0- 523 (17.4) 255 (8.5) 11.9 >12.0 43 (1.4) 47 (1.6) The calculation of odds ratio is based upon the formula (axc)/(bxd), where ‘a’ and ‘c’ are the number of subjects in each of the groups which are in agreement with the hypothesis. And ‘b’ and ‘d’ are the number subjects in each of the groups which are in disagreement with the hypothesis. Using this information calculate the following odds ratios for hypertension compared to individuals consuming less than 3g sodium per day. You may wish to draw a contingency table to help with your calculations. Risk of hypertension associated with sodium intake over 6 g/d Risk of hypertension associated with sodium intake over 9 g/d Risk of hypertension associated with sodium intake over 12g/d Meta-analysis 39. The figure below is from a systematic review and meta-analysis considering the impact of consuming foods rich in saturated fat upon risk of death from cancer. Considering the information presented in the forest plots, identify the correct statements from the list below. False True i. The strongest evidence of association between meat and cancer comes from the smallest studies. This suggests publication bias. ii. There is no association between milk intake and cancer mortality iii. High consumption of foods rich in saturated fat decreases risk of cancer death by 14% iv. High consumption of foods rich in saturated fat increases risk of cancer death by 14% v. Risk of cancer mortality associated with meat is greater than for processed meat vi. The lack of consensus in the results means that no conclusions can be drawn in relation to meat intake and risk of cancer death Standardized ratios 40. Note: There is negative marking for incorrect answers on this question The table below provides information on mortality due to oesophageal cancer in different regions of the UK. Region Crude mortality rate Standardized (deaths/100000) mortality ratio Wales 15.7 113.7 Scotland 16.5 119.6 SE England 8.4 60.9 Midlands 13.2 95.7 SW England 9.8 71.0 N England 14.1 102.2 Which regions have a higher mortality than the UK as a whole? SE England, SW England and Midlands Wales, Midlands and N England Scotland, Midlands and Wales Scotland, Wales and N England