Epi exam 3
We are interested in determining the average age of Type II diabetes diagnosis in Orlando, FL. We visit 2 hospitals and pull from the medical records the age of diagnosis for 100 people with diabetes. We determine the mean age at diagnosis to be 35.3 and find the standard deviation to be 5.65. What is the standard error (SEM)?
.565
Six thousand seven hundred fifty (6750) people who were free of disease X were enrolled in a cohort study in 1985 and followed with annual exams and interviews through 1995. Exposure to factor A was determined at study enrollment and the participants were followed until 1995 to observe new cases of disease X. Data from the study at the end of follow-up are shown in the following table: What is the incidence rate of disease X among persons exposed to factor A?
120/ 2880 = 0.04
What is the relative risk for the effect of exposure to factor A on disease X?
120/2880= 0.041 **** 30/3720= 0.008 ***** 0.04/0.008 = 5.00
Using the additive model (if no interaction is taking place) what is the attributable risk for both exposures?
20-10=10 *** 50-10=40 **** 40+10 = 50
Using the additive model, (if no interaction is taking place) what value would we expect in cell d?
20-10=10 *** 50-10=40 **** 40+10 = 50 **** baseline = 10 **** 10+50=60 [add baseline to attribute risk]
Using the multiplicative model what would be the expected relative risk for both exposures?
20/ 10 = 2 **** 50/ 10 = 5 **** 2 x 5 = 10
A prospective cohort study followed smokers and non-smokers to determine their development of coronary heart diseases (CHD). The table below represents the data collected from the cohort study. What is the absolute risk for developing CHD in smokers?
25 per 1000
By how much did participants reduce their caloric intake?
25%
What is the Odds Ratio for CHD in smokers to non-smokers?
250 X 9750 = 2437500 ***** 100 X 9900 = 990000 **** 2437500/990000 = 2.46 or 2.53
What is the relative risk of CHD in smokers to non-smokers?
250/ total = 250/ 10,000 = 0.025 **** 100/ total= 100/ 10,000= 0.01 ****** 0.025/ 0.01 = 2.5
We are interested in determining the average age of Type II diabetes diagnosis in Orlando, FL. We visit 2 hospitals and pull from the medical records the age of diagnosis for 100 people with diabetes. We determine the mean age at diagnosis to be 35.3 and find the standard deviation to be 5.65. What is the 95% CI for mean age of diabetes diagnosis?
34.19-36.41
Which percentage of the psychology studies failed to replicate?
64%
The journalist in the NPR story suggest that what type of study would be needed to determine if caloric restriction extends life?
A study that follows people for their life time
Why does the author refer to the p-value as a notoriously misleading statistic?
Because a p-value of .04 and .05 are in essentially the same thing, but one is significant and the other is not.
What is the goal of registered reports?
By registering the study parameters beforehand researchers are required to post finds without going back to do "p-hacking" You get feedback from peers before collecting data which helps improve the design of the study. To reduce reproducibility problems
In a study of 100 cases of colon cancer in women, there were 200 age- and race-matched controls. The suspected etiologic factor was higher-than-average consumption of red meat. The absolute risk of colon cancer in persons with this level of consumption is
Cannot be calculated
What is the correct sequence generally seen in the human studies:
Clinical Observations, Available Data, Case-control study, Cohort study, & Randomized Trials.
In a case-control study of obesity and adult-onset asthma, controls are matched to cases on the basis of race and gender. This approach to selection is intended to decrease the influence of which type of bias?
Confounding
U.S. Preventive Services Task Force is responsible for developing clinical practice guidelines for prevention and screening. What is the first and most important the USPSTF asks when making recommendations?
Does screening for X reduce morbidity and/or mortality?
Did the core body temperature in those that ate less go up or down.
Down
An example of misclassification bias is when scientists mislabel the study design they are using in their research.
False
Bias from "surrogate interviews" is a special case of bias that usually only occurs when dealing with surrogate pregnancies.
False
Confounding is a type of bias, which results in erroneous causal inferences.
False
Non-different misclassification most commonly shifts your RR or OR away from 1.
False
Positive interaction is called antagonism while the negative interaction is called synergism.
False
Randomized Control Studies are only ethical to use when the exposure is highly toxic or carcinogenic
False
The 3 studies that can be used to determine associations between variables are case report, cohort studies, and systematic review.
False
The degrees of freedom can generally be calculated as the sample size minus two.
False
The formula for calculating the OR when working with matched data is OR = c/b.
False
The two types of misclassification bias are information bias and selection bias.
False
When working with matched data in case-control studies you use the concordant data to calculate the odds ratio.
False
After Listening to Peter Donnelly, on average what combination do you reach on average more quickly "HTH" or HTT"?
HTT
Which of the following is an advantage of the case-control study design?
It is possible to study more than one exposure
The relative risk for developing oral cancer in an individual that drinks less than .4 oz of alcohol per day is .04. What does this RR say about the moderate consumption of alcohol?
It seems to prevent oral cancer.
What effect did the decrease caloric intake have on peoples metabolism?
It slowed down
In selecting the study population in case-control studies, controls are often selected by matching each one to a case according to variables that are known to be related to disease risk, such as sex, age, or race (individual matching or matched pairs). Given the table below, what is the matched pairs odds ratio?
Matched pairs OR = b/c ****** 1000/250 = 4
If a value of cell d were 60 what type of interaction (if any) would we be seeing?
No interaction
Suppose that one third of all cervical cancer cases were smokers as were one third of all controls and smoking status is independent of HPV infection. Is smoking a potential confounder in this study?
No, the chance of being a smoker among women with cervical cancer is the same as that of being a smoker among women without cervical cancer
In a case-control study of computer display exposure and glaucoma, cases and controls were also asked about television watching habits. Errors in recall of exposure to video screens occurred with equal frequency among cases and controls. Which one of the following biases likely occurred?
Non-differential misclassification
The US Public Health Service Stage I guidelines for making causal inferences defines studies by their quality. Which of the following is the HIGHEST quality study?
Randomized, double-blind, placebo-controlled.
Which of the following is a type of information bias?
Reporting bias
Which of the following in NOT one of the 3 reasons given for why the studies failed to replicate?
Researchers who conducted the replication studies failed to asked the original authors to scrutinize the replication plan and did not ask them to provide feedback.
if we are studying the possible relationship ofbenzene exposure and lung cancer and the response rate of potential subjects is higher in people with lung cancer who were exposed to benzene than in people with the lung cancer who were not exposed, an apparent association could be observed even if in reality there is no association. The following is an example of what type of error in epidemiology?
Selection Bias
Which of the following is the LEAST important of the guidelines suggested by the Surgeon General in determining the etiology of chronic diseases?
Specificity of Association
The author states that science has a "false positive problem". What is she implying?
That there have been several high-profile failures to repeat the results of classic experiments
Which of the following is the logical error the Peter Donnelly ascribed to the pediatrician's reasoning in the odds of SIDS in two children?
The pediatrician assumed the events were independent
Why are controls needed in a case-control study?
They do not have the outcome of interest and allow for estimation of exposure frequency among those without the disease
At the end of a cohort study, one establishes 4 groups. These groups are exposed with no disease, exposed with disease, non-exposed with no disease, and non-exposed with disease
True
Confounding is not an error in the study, but rather is a true phenomenon that is identified in a study and must be understood.
True
Discovering a confounding variable can be useful in identifying groups that are at high risk for a particular disease.
True
Exclusion bias is a type of selection bias.
True
In Type I error, we reject the null hypothesis when it is true, and conclude that there is an effect when, in reality, there is none.
True
In Type II error, we do not reject the null hypothesis when it is false, and conclude that there is no effect when one really exists
True
Most experts accept any effect greater than additive as evidence of positive interaction, which is also called synergism
True
Odds Ratios = The Cross Products ratio.
True
Odds ratios are the most appropriate measure of association for Case-Control Studies.
True
Relative risk is the most appropriate measure for Cohort studies
True
The algorithm for determining interaction consists of the three following questions: 1.) Is there an association? 2.) If so, is it due to confounding? 3.) Is there an association equally strong in strata formed on the basis of a third variable?
True
There are four different methods for addressing confounding, these include; individual matching, group matching, adjustment, and stratification.
True
Three important factors that negatively affect our ability to make causal inferences are, bias, confounding, and interaction.
True
Use the table below to check the veracity of the statement. The formula for calculating RR is (a/a+b) / (c/c+d)
True
Use the table below to check the veracity of the statement. The formula for determining absolute risk of the disease in the exposed group is (a/a+b)
True
Using the additive model for determining interaction. The value in cell "d" should be 20. ( The numbers in the table represent incidence rates)
True
Using the multiplicative model for determining interaction. The value in cell "d" should be 30. ( The numbers in the table represent incidence rates)
True
We estimate the value of a parameter using the data collected from a sample.
True
When calculating a 95%confidence interval, 1.96 represents the number of standard deviations from the mean
True
When the sample size is large, the difference between the t-distributions and the normal distribution is negligible.
True
When we only have one sample, our best estimate of the population mean is the sample mean.
True
When working with unmatched data in a case-control study the proper formula for calculating the odds ratios is OR = ad/bc.
True
Where the data are not normally distributed, and/or we do not know the population variance but estimate it by s2, the sample mean follows a t-distribution.
True
Matching is employed in a case-control study in order to ensure that
Variables associated with the outcome under study are controlled for in both the case and control groups
We are interested in determining the average age of Type II diabetes diagnosis in Orlando, FL. We visit 2 hospitals and pull from the medical records the age of diagnosis for 100 people with diabetes. We determine the mean age at diagnosis to be 35.3 and find the standard deviation to be 5.65. After determining the CI which of the following is a TRUE statement that can be made regarding the CI.
We are 95% certain that the true mean age of diabetes diagnosis in the population lies in the CI we have calculated.
What is Type II error?
We do not reject the null hypothesis when it is false, and conclude that there is no effect when one really exists
Which of the following is Peter Donnelly's final point about uncertainty?
We get question of uncertainty wrong all the time.
What is Type I error?
We reject the null hypothesis when it is true, and conclude that there is an effect when, in reality, there is none.
Is the OR in this study a good approximation of the RR?
Yes
Using the multiplicative model, what value would we expect in cell d?
[relative risk] x [base line] = 10 x 10 = 100
Which of the following are TRUE as it pertains to interpreting relative risk?
a. If the relative risk is equal to 1, the risk in exposed persons equals the risk in non-exposed persons. b. If the relative risk is greater than 1, the risk in exposed persons is greater than the risk in non-exposed persons. c. If the relative risk is less than 1, the risk in exposed persons is less than the risk in non - exposed persons. ALL OF THE ABOVE
Which of the following methods can be used to address confounding in case-control studies?
a. Individuals Matching. b. Stratification c. Group Matching d. All of the above.
When is odds ratio obtained in a case-control study a reliable approximation of the relative risk for the general population?
a. The exposure distribution among cases is representative of all persons with disease b. The exposure opportunity for controls is equivalent to the distribution of exposure opportunities for the population without the disease c .When the disease under investigation is rare in the population d. All of the above
By which of the biological mechanisms below would a reduced diet extend life?
by reducing oxidants
A convenience sample is superior to a random sample because any resulting conclusions will be more generalizable to the population
f
In case-control studies, we can either use relative risk or odds ratios to determine the association between the exposure and the disease.
false
Use the table below to check the veracity of the statement. The OR for this study is 2.
false
According to his wife what does Peter Donnelly model?
genes
Which of the following are TRUE regarding p-values? i. The P-values represents the probability of obtaining our results, or something more extreme, if the null hypothesis is true. ii. The smaller the P-value, the greater the evidence against the null hypothesis. iii. The smaller the P-value, the greater the evidence against the alternative hypothesis
i & ii only
Which of the following are appropriate measures of association when interpreting cohort studies? i. Relative Risk ii. Odds Ratio iii. Prevalence
i & ii only
When is the OR a good estimate of the RR? i. When thecasesstudied are representative, with regard to history of exposure, of all people with the disease in the population from which the cases were drawn. ii. When thecontrolsstudied are representative, with regard to history of exposure, of all people without the disease in the population from which the cases were drawn. iii. When the disease being studied does not occur frequentlya.
i, ii, & iii only
Which of the following are TRUE with regards to discordant pairs? i. They are pairs in which both the case and control were exposed. ii. They are pairs in which the case was exposed but the control was not exposed. iii. They are pairs in which the control was exposed but the case was not exposed.
ii & iii only
What is the relative risk for developing oral cancer in cigarette smokers that smoke greater or equal to 40 packs of cigarettes per day?
relative risk = division 2.34 / 1.00 = 2.43
When we take a sample of a population we introduce the potential for error into any conclusions that might be drawn about that population
t
What is an odds ratio?
the ratio of the odds of developing the disease in exposed persons to the odds of developing of disease in non-exposed persons.
Use the table below to check the veracity of the statement. The formula for calculating OR is ad/bc.
true
Use the table below to check the veracity of the statement. The formula to determine the odds of an exposed person developing the disease is a/b.
true
If 100 in a 1,000,000 people have HIV what number of false positives would we see if we tested all 1,000,000 people? ( Note: Assume the test is 99% accurate)
~10,000
