A. Some Basics (Chapter 1 )
1. As you know, epidemiology is the study of the distribution and determinants of disease frequency in humans and the application of this study to control health problems. Briefly describe two different types of activities that epidemiologists perform.
2. A group of researchers met to plan a study of causes of poor eating habits among graduate students in public health. After the meeting, the group left their building, and walked outside into a raging snowstorm. Their planning notes were ripped from their hands, and tossed to the street. Although the notes were recovered, they were hopelessly mangled and out of order. Suppose that you were the team member who offered to re-construct the notes. Describe in a logical order at least three steps that the team should undertake to develop, test, and evaluate a hypothesis on the causes of poor eating habits among public health students. Be sure to state what type of study design you would use and why.
B. Outbreaks of Gastroenteritis on Cruise Ships (Chapters 2, 3)
Recently, a report in the Centers for Disease Control (CDC) publication Morbidity and Mortality Weekly described a string of individual cases of Norwalk virus gastroenteritis occurring on cruise ships. From this report, the editor went on to form a number of hypotheses as to why there has been and this rather unusual increase in reported gastroenteritis outbreaks on cruise ships in 2002.
1. What type of study is being described in this report?
2. What is the main limitation of this type of study?
The cruise ship owners contacted the CDC to conduct an in-depth analysis of the possible modes of transmission of the Norwalk virus in the cruise ship environment. CDC investigators interviewed all of the passengers on the last affected cruise (N=3,000) and obtained information on the passenger’s recreational activities. They found the following results: 1,000 passengers had gone swimming in the upper deck pool and 2,000 passengers had never gone swimming in the upper deck pool. 100 of the passengers who swam in the upper deck pool and 100 of the passengers who did not swim in this pool developed Norwalk virus gastroenteritis during the cruise. FYI: The cruise lasted one week.
3. Set up the two by two table for these data.
4. Calculate the risk ratio of gastroenteritis associated with swimming in the upper deck pool.
5. State in words your interpretation of the above risk ratio.
6. Calculate the risk difference in the above example.
7. State in words your interpretation of the above risk difference.
8. What measure of disease frequency are you using to calculate the risk ratio and risk difference?
9. What measure of association should be used to answer the question, “How many additional cases of Norwalk virus gastroenteritis among all cruise passengers (N=3,000) was associated with swimming in the upper deck pool?”
10. When the cruise ship owners examined the findings of the CDC investigation, they stated that the crude results (calculated above) were invalid because of the age differences between the people who swam in the upper deck pool and those who did not. Examine the following table and state whether or not you agree with the cruise ship owners.
Age Groups (years) Swam in Upper Deck Pool Never Swam in Upper Deck Pool
10-20 25% 10%
21-30 20% 10%
31-40 20% 15%
41-50 15% 20%
51-60 10% 20%
>60 10% 25%
Total 100% 100%
a. Based on these data, do you agree or disagree with the assessment that the crude results were invalid? Briefly justify your answer.
b. Regardless of whether you agree with the cruise ship owners, explain in at least 3 sentences the method that epidemiologists use to account for age differences in populations. Be sure to mention what additional data would be needed to perform this procedure.
C. Tuberculosis in a Housing Community (Chapter 2)
In January 2000 you began a one-year study of tuberculosis (TB) in a subsidized housing community in the Lower East Side of New York City. You enrolled 500 residents in your study and checked on their TB status on a monthly basis. At the start of your study on January 1st, you screened all 500 residents. Upon screening, you found that 20 of the healthy residents were immigrants who were vaccinated for TB and so were not at risk. Another 30 residents already had existing cases of TB on January 1st. On February 1st, 5 residents developed TB. On April 1st, 5 more residents developed TB. On June 1st, 10 healthy residents moved away from New York City were lost to follow-up. On July 1st, 10 of the residents who had existing TB on January 1st died from their disease. The study ended on December 31, 2000. Assume that once a person gets TB, they have it for the duration of the study, and assume that all remaining residents stayed healthy and were not lost to follow-up.
1. Is the subsidized housing community in the Lower East Side of New York City a dynamic or fixed population? Briefly explain the rationale for your answer.
2. What was the prevalence of TB in the screened community on January 1st ?
3. What was the prevalence of TB on June 30th ?
4. What was the cumulative incidence of TB over the year?
5. Suppose that you wanted to calculate the incidence rate of TB in the study population. Calculate the amount of person-time that would go in the denominator of this incidence rate. Be sure to show your work.
6. What was the case-fatality rate among residents with TB over the course of the year?
D. Measures of Disease Frequency and Association (Chapters 2, 3)
1. Consider a group of 1,000 newborn infants. 100 infants were born with serious birth defects and 20 of these 100 died during the first year of life. 90 of the 900 remaining infants without any birth defects also died during the first year of life.
a. Calculate the prevalence of serious defects in this population at the time of birth.
b. Calculate the overall cumulative incidence of mortality in this population.
c. Calculate the cumulative incidence difference in mortality between infants born with serious birth defects and without.
d. State in words your interpretation of the cumulative incidence difference calculated in part c.
2. How does each of the following conditions influence the prevalence of a disease in a population? For each scenario, assume that no other changes occur. Your choices are: increases prevalence, decreases prevalence, or has no effect on prevalence.
a. A treatment is developed that prolongs the life of people suffering from the disease
b. A new measure is developed that prevents new cases of disease from occurring
c. There is immigration of a large number of healthy people into the population.
3. In January, 1999 forty heterosexual hemophiliac patients (all males) who periodically received intravenous infusion of blood products to control their hemophilia were asked to participate in a 3 year prospective study to determine their risk of three adverse outcomes: a) the rate of HIV (human immunodeficiency virus) seropositivity, b) the rate of developing clinical signs of AIDS, and c) death from AIDS. The men were to undergo an interview, physical examination, and blood testing every 6 months for 3 years.
Among the 40 subjects there were 30 who were seronegative and healthy for the entire duration of the study, and all of these were followed for the entire 3 years.
During the initial screening, 2 of the men were found to already be HIV+, although none of them had clinical signs of AIDS. The table below describes the 10 subjects who either developed AIDS, or became lost to follow-up, or already had AIDS at the start of the study.
HIV+ = found to be HIV+ at the very beginning of the interval observation period
? = lost to follow-up
************* Follow-up **********************
Screening Jan. 1999 June 1999 Jan. 2000 June 2000 Jan. 2001 June 2001
1 --------- --------- --------- --------- --------- HIV+--
2 --------- --------- --------- --------- --------- ?
3 --------- --------- HIV+-- --------- --------- ---------
4 --------- ?
5 --------- --------- --------- ?
6 --------- --------- --------- --------- --------- ?
7 --------- HIV+-- --------- --------- --------- ---------
8 --------- --------- --------- --------- --------- ?
9 HIV+ --------- --------- --------- --------- --------- ---------
10 HIV+ --------- --------- --------- --------- --------- ?
a. From the information above, what was the prevalence of seropositivity (HIV+) in Jan. 2000?
b. What was the cumulative incidence of seropositivity (HIV+) during the 3 year study?
c. What was the incidence rate of seropositivity (HIV+) during the study? (Show your work.)
4. If an exposure has no association with a certain outcome, then what value would you obtain for each of the measures of association listed below? Fill in all the blanks.
A. Rate ratio ________________
B. Odds ratio ________________
C. Rate difference _____________
D. Attributable proportion in the exposed___________
5. A group of 100 healthy women was followed prospectively for 10 years. All subjects entered the study on January 1, 1990 and all women were followed until December 31, 1999. None were lost to follow-up. During this period, 5 subjects were diagnosed with breast cancer, but they all survived to the end of the study. The time at which these 5 subjects developed cancer is shown in this table. Assume that each diagnosis occurred exactly half way through the year.
Subject 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999
a. What was the cumulative incidence of breast cancer?
b. What was the incidence rate of breast cancer?
c. What was the prevalence of breast cancer “survivors” on December 31, 1999?
6. Quote from a news report: “Based on interviews and blood samples of over 8,000 persons, 11% of persons classified as heavy drinkers of alcohol were involved in home accidents over a three-month period.” Can we conclude that there is an association between drinking and home accidents from these data? Circle YES or NO and briefly justify your answer.
E. Study Designs (Chapters 6-9)
1. Briefly define each of the following terms associated with experimental studies. Do not simply repeat the words given in the term.
a. Single-masked study (also called single-blinded study)
c. Reference population
2. a. Suppose that an experimental study investigator randomized 100 men to the treatment group and 100 men to the comparison group. However, only 75 men in the treatment group and 85 men in the comparison group were able to comply exactly with their assigned regimens. Should the data analysis be based on the 200 men originally randomized or the 160 men who were able to comply? Give the reasons for your answer.
3. State an important difference and an important similarity between experimental and cohort studies.
4. A cross-sectional study was conducted on the association between passive smoke inhalation and the occurrence of dental caries in children. (Passive smoke exposure occurs when children live with family members who smoke.) The investigators thought that conclusions from this study were limited because of the cross-sectional nature of the data. Suppose that they asked you for advice and you told them that they should have conducted a prospective cohort study because it is a better study design.
a. Briefly describe the specific limitation of a cross-sectional study is avoided by conducting a prospective cohort study.
b. The investigators take your advice and hire you to help them design a new prospective cohort study. They want to know if they should use a special or a general cohort to assemble the exposed population. Briefly describe each of these options, tell them which one is best in this setting, and explain your reasons.
Which option is best in this setting?
c. The investigators also ask you about the options for selecting a comparison group in a cohort study. Briefly describe the three different options and state which one is best in this setting and why.
d. The investigators want to evaluate the children’s exposure to passive smoke by measuring serum cotinine levels. (Serum cotinine is a biological marker of recent tobacco smoke exposure that involves performing a laboratory test on a blood sample.) Describe one advantage and one disadvantage of using this method for ascertaining this exposure.
5. Phthalates are present in many diverse products, including insect repellents, body lotions, perfumes, and food packaging. Because animal experiments suggest that phthalates may have an adverse effect on the male reproductive system, a group of infertility specialists decided to conduct a case-control study on phthalate exposure and sperm abnormalities in adult men.100 cases with sperm abnormalities and 100 controls were identified and enrolled from among patients at their infertility clinic. 30 cases and 10 controls had high urinary phthalate levels; the remainder had normal urinary phthalate levels.
a. Set up the two by two table using these data.
b. Use these data to calculate the odds ratio describing the relationship between phthalate levels and sperm abnormalities. (5 pts)
c. State in words your interpretation of this odds ratio.
d. The infertility specialists never took any epidemiology courses and so they do not know the basic purpose of the control group in a case-control study. What would you tell them?
e. Despite their lack of knowledge, the specialists thought that they selected an appropriate control group for their study. The control group included men who were evaluated at the clinic for infertility but who turned out to have normal sperm parameters. (The female partner of the controls turned out to be the source of infertility.) Do you agree or disagree with the specialists’ opinion that this is an appropriate control group for this study? State the reason for your answer.
6. Which type of observational study design is best suited for each of the following scenarios? Your choices are: retrospective cohort, prospective cohort, and case-control study. If there is more than one answer, then give them all.
a. Little is known about the causes of the disease
b. The exposure under study is rare
c. The disease has a long latent and induction period
7. What specific advantage does randomization have as compared to all other methods used to control for confounding? In other words, what can randomization do that no other method to control confounding can do?
8. A study was done to determine whether the amount of money spent on soft drinks was related to mortality from diabetes. The investigators collected data on per capita (average per person) soft drink consumption in ten US states and examined its relationship to mortality rates from diabetes in those ten states. In order to calculate per capita sales they gathered annual data on soft drink sales from commerce records and then divided these figures by the state’s population from the most recent census. The mortality data were gathered from the vital records department in each state. Here are the data that they collected.
US State Annual per Capita Soft Drink Sales Annual Diabetes Mortality Rate (per 100,000 population)
Massachusetts $150 207
New York $300 353
Florida $500 688
Alabama $700 801
Alaska $50 75
California $500 605
Nevada $200 310
Idaho $250 325
Ohio $400 454
Arkansas $350 405
a. What type of study is this?
b. Unfortunately, the investigators had never taken and epidemiology course and did not acknowledge any limitations of their study. Briefly describe two major limitations that they should have acknowledged.
F. Obesity and Cancer Risk (Chapters 2, 3, 6, 10-12, 15)
A recent study followed 900,000 US adults from 1982 to 1998. At baseline, all participants were screened and determined to be cancer free and their body mass index (BMI) was calculated. Body mass index is a measure of obesity that is calculated using a person’s height and weight. Subjects were separated into the following groups according to their BMI: (a) normal weight, (b) slightly overweight, (c) moderately overweight and (d) greatly overweight. 57,145 deaths from cancer occurred in the population during the follow-up period.
1. What type of study is this?
2.. Use the data given above to calculate the cumulative incidence of deaths from cancer among the study population over the follow-up period.
3. What additional information would need to be provided for you to be able to calculate the incidence rate of cancer deaths?
4. The following results were seen for men and women when the heaviest members of the cohort were compared to those with normal weight:
a. State in words your interpretation of the risk ratio given for the men.
b. State in words your interpretation of the 95% confidence interval given for men. (Do not merely use the confidence interval to assess statistical significance.)
c. Are these results confounded by gender?
5. The authors stated that they controlled for confounding many risk factors using a multivariate analysis. State an alternative method that the authors could have used to control for confounding in the design or analysis. In addition, name two confounding variables that you think should be controlled using this method.
6. To show the public health impact of obesity on cancer mortality in the total population, the authors stated that 90,000 cancer deaths could be prevented each year in the total US population if adults could maintain normal weight. What measure of association is being described by the authors?
7. This study used self-administered mail questionnaires to gather data on height and weight in order to calculate the measure of obesity. Which of the following types of problems were SURELY AVOIDED by this method of data collection.
1) Interviewer Bias
2) Exposure misclassification
4) Selection Bias
8. Based on these sufficient/component causal model, indicate whether each of the following statements is true or false.
a. Being greatly overweight is a necessary cause of cancer mortality.
b. Being greatly overweight is a component cause of cancer mortality.
G. Cell Phone Use and Brain Cancer (Chapter 9, 11, 12, 13)
The association between cellular telephone use and the risk of brain cancer was investigated in a case-control study. The study included 475 cases and 400 controls and the following results were seen:
Cellular Yes 270 200
User No 205 200
Total 475 400
1. Calculate the odds ratio based on these data.
2. The p value for this odds ratio is 0.06. State your interpretation of this p-value.
3. Gender was considered a potential confounder and effect measure modifier in this study. The data were stratified into males and females in order to assess these issues.
Cases Controls Cases Controls
Cell Yes 242 150 Yes 28 50
Users No 100 50 No 105 150
Stratum-specific OR = 0.8 Stratum-specific OR = 0.8
a. Is gender a confounder in this study?
b. Briefly justify your answer to part a.
c. Is gender an effect measure modifier in this study?
d. Briefly justify your answer to part c.
H. Optimism and Cancer Survival (Chapter 2, 3)
1. Recently, Australian researchers conducted a study of the relationship between optimism and colon cancer survival. Their hypothesis was that colon cancer patients who had a positive outlook on life would have a lower five-year cumulative incidence of mortality. The study included 100 recently diagnosed colon cancer patients who underwent psychological testing and were found to have a optimistic outlook on life and 100 recently diagnosed colon cancer patients who underwent the same psychological tests and were found to have a pessimistic outlook on life. By the end of five years of follow-up, 50 of the 100 patients with the optimistic outlook and 75 of the 100 patients with the pessimistic outlook had died from colon cancer.
a. Set up and fill in the two by two table using these data.
b. What is the prevalence of colon cancer in the study population?
c. Compare the cumulative incidence of mortality in the optimistic group to the cumulative incidence of mortality in the pessimistic group using a ratio measure of association.
d. State in words your interpretation of the result you found in part c.
I. More on Confounding (Chapter 11)
1. Consider each of the following scenarios and state whether or not the variable in question is a confounder.
A study of the risk of pulmonary hypertension among women who take diet drugs to lose weight. The crude relative risk of pulmonary hypertension comparing diet drug users to non-users is 17.0 and the age adjusted relative risk is 5.0. Is age a confounder in this study?
A cohort study of liver cancer among alcoholics. Incidence rates of liver cancer among alcoholic men are compared to a group of non-alcolohic men. Is gender a confounder in this study?
A case-control study of the risk of beer consumption and oral cancer among men. In this
study, cigarette smoking is associated with beer consumption and is a risk factor for oral
cancer among both beer drinkers and nondrinkers. Is cigarette smoking a confounder in this
J. Screening for Disease Control (Chapter 16)
1. Below is a chart of the natural history of disease for one individual.
A B C D E F
A = Biological onset
B = Disease detectable by screening
C = Individual’s disease was detected by screening, the diagnosis was made, and treatment was begun.
D = Clinical symptoms would have developed if the individual was not screened
E = Death would have occurred due to the disease if the individual had not been screened and treated
F = Death actually occurred
Give letters in the diagram associated with the start and stopping points for each of the following.
a. The detectable pre-clinical phase for this individual
b. The lead time for this individual
c. Increased survival time that the screening gave this individual
2. Suppose that your company has just developed a new screening test for a disease and you are in charge of testing its validity and feasibility. You decide to evaluate the test on 1000 individuals and compare the results of the new test to the gold standard. You know the prevalence of disease in your population is 30%. The screening test gave a positive result for 292 individuals. 285 of these individuals actually had the disease on the basis of the gold standard determination.
a) Fill in all cells of the two by two table.
Gold Standard Determination of Disease
Results of Screening Test Yes No
b. Calculate the sensitivity of the new screening test.
c. Interpret the results of sensitivity calculation in one sentence.
d. Calculate the specificity of the new screening test.
e. Interpret the results of specificity calculation in one sentence.
f. Calculate the predictive value of a positive test.
g. Interpret the results of the predictive value positive calculation in one sentence.
h. What would happen to the predictive value positive if this test were administered in a population with a disease prevalence of 1% instead of 30%? (Note that the sensitivity and specificity of the test remain the same.) Would it remain the same? Increase? Decrease?
L. Making Fair Comparisons (Chapter 3)
1. Suppose there are 10,000 people in City A and 10,000 people in City B. The following data give the age distribution of the two cities and the crude and age-specific cumulative incidences (CI) of influenza in the two cities during 2004.
City A City B
Total population 10,000 10,000
Number of “young” people 4,000 6,000
Number of “old” people 6,000 4,000
Crude CI 4.28% 2.92%
CI among “young” people 0.2% 0.2%
CI among “old” people 7.0% 7.0%
a. Is age-standardization needed to make a fair comparison between the crude cumulative incidences in Cities A and B?
b. Write one-two sentences to justify your answer. (3 points)
2. A study compared Miami, Florida and Boston, Massachusetts with respect to cancer mortality between the ages of 18 and 99 years.
In Miami, the crude rate for death from cancer was 220 per 10,000 population per year.
In Boston, the crude rate for death from cancer was 150 per 10,000 population per year.
One concern raised by this study was the possible effect of different age distributions in the two cities. You check on the age distributions and find that Boston’s population had a much greater percentage of young people (college students and young professionals), while Miami had a larger percentage of people in the older age brackets (retired persons).
Because of the difference in age distribution, the authors of the study used direct standardization to calculate age-adjusted rates using the age distribution for the US population in 2000 as the standard.
In Miami, the age-adjusted rate for death from cancer was 108 per 10,000 population per year
In Boston, the age-adjusted rate for death from cancer was 175 per 10,000 population per year
(a) Did the differences in age contribute to the observed difference in the crude rates ? Why or why not?
(b) Your grandfather who is 65 years old and lives in Ohio is about to retire and move to Miami. However, his friend advised him not to move these because “A lot of folks get cancer down there.” So now, he is considering moving to Boston instead. Based on these data, what advice would you give him about the risk of dying from cancer in Miami as compared to Boston?
N. Random Error (Chapter 12)
1. A randomized experimental study was conducted to evaluate the effectiveness of a new pertussis vaccine. One-thousand healthy children were randomized to receive either the new vaccine (500 children) or the old vaccine (500 children). The children were followed for two years to monitor the incidence of pertussis. At the end of the study, the risk ratio for developing pertussis was 0.5 among the children who received the new vaccine compared to children who received the old vaccine. The 95% confidence interval for this relative risk was 0.2-0.8 and the p value was 0.01.
a. State in words your interpretation of the risk ratio. Be as descriptive as possible.
b. State in words your interpretation of the p value. Be as descriptive as possible.
c. State in words your interpretation of the 95% confidence interval. Be as descriptive as possible, and do not simply repeat the interpretation give in part b.
O. Causation (Chapter 15)
1. Use your knowledge of the sufficient-component causal model to fill in the blanks with one of the following terms: sufficient cause, component cause, necessary cause.
a. ……………..can act far apart in time
b. The completion of a ………… is synonymous with the occurrence
c. Blocking the action of a ………… will prevent all cases of disease
by all causal mechanisms.
d. Every sufficient cause always has at least two ……………..
2. Which two study designs provide the best evidence to support Hill’s causal guideline on temporality? Be as specific as possible.
1. Experimental study
2. Prospective cohort study
3. Describe in 1-2 sentences Hill’s reason(s) for suggesting that strong associations are more likely to be causal than weak associations. (2 points)
P. Stratification and Effect Measure Modification (Chapter 13)
1. A cohort study was undertaken to examine the association between high lipid level and coronary heart disease. Participants were classified as having either a high lipid level (exposed) or a low or normal lipid level (unexposed). Because age is associated with both lipid level and risk of heart disease, age was considered a potential confounder and the age of each subject was recorded. The following data describes the study participants:
Developed CHD No CHD
EXPOSED (High lipid level)
young 20 3980
old 200 5800
UNEXPOSED (Low or normal lipid level)
young 18 6982
old 65 2935
a. Calculate the appropriate crude ratio measure of association combining the data for young and old individuals.
b. For the same cohort study, perform a stratified analysis and calculate the appropriate stratum-specific ratio measures of association. What are they?
c. Do the data provide evidence of effect measure modification on the ratio scale? Justify your answer.
Dikerjakan 2 minggu dari sekarang 14 Januari 2009 samapai 1 Februari 2009
Selamat mengerjakan...... semoga sukses