Ep101.1-4 Fa_2010.11_v1.0.pdf

EP101

Fundamentals of epidemiology

Please note: 

FA 1 should be completed after studying sessions 1 to 11 of Fundamentals of Epidemiology;



FA 2 should be completed after studying session 10 of Fundamentals of Epidemiology;



FAs 3 and 4 should be completed after studying sessions 11 to 17 of Fundamentals of Epidemiology.

FINAL DEADLINE FOR SUBMISSION VIA AMS: 31 March

EP101 Formative Assignment 1 In a cohort study of upper respiratory tract infections (URTI) in 551 infants, 445 infants were exclusively bottle-fed and 106 infants were exclusively breast-fed. 207 of the bottle-fed babies, and 34 of the breast-fed babies had at least one URTI in the first four months of life.

1. Draw up a summary table to represent this data.

(10%)

2. Calculate the risk ratio of one or more episodes of URTI in bottle-fed compared to breast-fed infants and interpret this briefly. Comment on any other measures of effect that could also be calculated using the data from this study. (30%)

3. Using a prevalence of bottle feeding in the general population at one month of age of 30%, calculate the population attributable fraction (PAF) of bottle feeding for URTI. What is the meaning of this fraction (in words)? What assumptions have been made in calculating it? (20%)

4. What do you conclude about the role of infant feeding and URTIs from this study? Can you think of non-causal reasons why there might be an association between infant feeding and URTI? What further evidence would you like to assess these? (20%)

5. What other issues would you need to consider before deciding on whether to promote a campaign to encourage breast-feeding? (20%)

EP101 Formative Assignment 2 The tables below give the population sizes (in thousands) of two standard regions in England and Wales and the number of deaths from Chronic Obstructive Pulmonary Disease (COPD) by age-group, from 35 to 84 years of age for 1990. The World Standard Population age distribution (in thousands) is also given1.

Age group 35–44 45–54 55–64 65–74 75–84

East Anglia Population (in 1000s) 137.6 114.5 102.7 89.0 47.0

Deaths 0 10 36 155 261

North West Population (in 1000s) 443.8 364.2 315.9 242.1 118.4

Deaths 9 53 267 841 1058

World Standard Population (in 1000s) 15 8 4 2 0.8

Please show all calculations. You might like to use a spread sheet package to do the calculations and then paste the table, intermediate and final results into a word processing document. 1. Calculate the crude death rates for age 35–84 from COPD for each of East Anglia and the North West regions. Compare the two rates. Comment on the age distribution of the two populations (showing any calculations you may do) and suggest why it may be inappropriate to compare two crude rates. (25%) 2. Using the World Standard population distribution calculate directly standardised rates for both East Anglia and the North West region over the age range 35–84. Comment on the difference between these standardised rates and the crude rates. (30%) 3. Calculate a Comparative Mortality Figure (CMF) for COPD mortality in East Anglia relative to the North West region. What does this show? (20%) 4. What information would you need in order to compare these populations using the indirect standardisation method? Do you have enough information to calculate standardised mortality ratios? (Only state the necessary information, no need to do the calculations here.) (5%) 5. Discuss the advantages and disadvantages of using direct standardisation compared to indirect standardisation, in the context of this example. (10%) 6. What general conclusion do you come to in relation to COPD mortality in these regions? (10%)

1Please note that here the World Standard Population is presented in 1000s. However, a standard population can be presented in 1,000s, 10,000s, 100,000s or millions, it does not matter because the numbers are only used to weight the estimates we obtain from the population of interest. In fact, some standard populations are presented as simple proportions of 1.00.

EP101 Formative Assignment 3 Autism is a psychiatric disease usually first recognised around age 2 years and diagnosed around age 4 or 5 years by a child psychiatrist. The initial measles vaccine is usually given with the mumps and rubella vaccine (MMR) at around 15 months of age. There was a recent claim, widely discussed in newspapers in England, that measles vaccine causes autism. You are asked to conduct a case-control study to test this hypothesis. (Please try and limit your answer to 1000 words.) 1. Who would be the cases and controls in your study? Discuss your criteria.

(25%)

2. What are the exposure and outcome of interest?

(10%)

3. What are the possible confounders you would consider in your study (justify your choice)? (16%) 4. Discuss the advantages and disadvantages of selecting your cases and controls from the following sources (these are sources which include all your cases): a) Children registered in primary health facilities i.e. controls from the same primary health facility as cases at the time the final diagnosis of autism was made. (12%) b) Children registered in psychiatric facilities i.e. controls from the same psychiatric facility as cases at the time the final diagnosis of autism was made. (12%) 5. Should you use mothers' information on the child's vaccination for your measure of exposure or try to find out vaccine status from a vaccination card? Explain the advantages and disadvantages of both sources of information. What would be the effect on the association between vaccine and autism? (25%)

EP101 Formative Assignment 4 A prospective cohort study of postmenopausal hormone replacement therapy (HRT) and primary prevention of cardiovascular disease was performed. The sample consisted of 70 533 nurses who completed a postal questionnaire in 1976 and were followed up until 1996 using biennial questionnaires. Participants were asked to report all nonfatal infarctions, confirmed by medical records where possible. Deaths (fatal coronary disease) were reported by participants’ families or identified via the National Death Index, with deaths considered to be due to coronary disease if medical records or autopsy findings (where available) confirmed a fatal myocardial infarction. 1258 major coronary events (nonfatal myocardial infarction or fatal coronary disease) were recorded during the follow-up period. Questionnaires collected information on HRT use and other relevant variables. Approximately 72% of eligible women agreed to take part in the study, and more than 90% of those taking part were successfully followed-up. The risk of major coronary events by HRT use was examined. The main findings are summarised in the table below: HRT use

Never Current

Personyears of follow-up 358 125 265 203

Number of Cases

Age adjusted Rate Ratio

Multivariate Rate Ratio*

662 259

1.00 0.54

1.00 0.61

*The following variables were adjusted for: age, body mass index, history of diabetes, hypertension, high cholesterol level, age at menopause, cigarette smoking, and parental history of premature heart disease.

1. Describe briefly what is meant by selection bias and discuss how it might be operating in this cohort study. (25%)

2. What do we mean by generalisability of results? Does this affect our interpretation of this study? (15%)

3. Does this study support the hypothesis that HRT reduces the risk of major coronary events? Explain your answer. What other possible explanations (other than selection bias) might there be for these results? (35%)

4. Controversy still remains regarding the possible association between HRT use and coronary heart disease in women. Briefly suggest an alternative study design that might provide further unconfounded evidence for the relationship between HRT and coronary heart disease, and discuss one strength and one limitation of this design. (25%)