HSC4501: EXAM 3 STUDY GUIDE, iCLICKERS, AND HW ASSIGNMENTS

The crude association between occupational exposure and kidney cancer (outcome) among the study sample was 3.2. When data were stratified by gender, the associations for female and male workers were 1.21 and 1.18 respectively. Assume that all (crude, and stratified) measures of association were statistically significant. Gender was (use 15% difference rule):

A confounder

Which of the following is (are) the source of selection bias in case-control studies? (select all that apply)

- hospital controls - prevalent cases

Which principle for guiding research involving human participants states that vulnerable populations who are incapable of providing informed consent are not to be used in research?

Justice

Which of the following is a limitation to using random assignment in a clinical trial?

May interfere with the doctor-patient relationship

If depression (exposure) status is measured with the same degree of error or inaccuracy for the diseased (suicidal attempt +) and non-diseased (suicidal attempt -) groups, it is called:

Non-differential or random misclassification of exposure

In a case-control study, a researcher found no association between exposure to A and disease B. Later she found out that there was a misclassification when measuring exposure to A, in both cases and controls. What is a potential explanation for the results (no association) of this case-control study?

Nondifferential classification

Which of the following best defines external validity?

The extent to which the results of a study are relevant to people who are not part of the study

Which of the following is an effective way to minimize the threat of observation bias (information bias)?

1. Proper training and blinding of interviewers as to case-control status 2. Use medical records or other recorded health information, rather than recalled information about exposure 3. Those from which information is being obtained should be unaware of the specific hypotheses being tested

The crude association between occupational exposure and lung cancer (outcome) among the study sample was 8.4. When data were stratified by alcohol intake, the associations were 2.2 (nondrinkers) and 14.5 (drinkers). Assume that all (crude, and stratified) measures of association were statistically significant. Alcohol use was (use 15% difference rule): Group of answer choices

An effect modifier

A (An) _______________ is a variable that differentially (positively and negatively) modifies the observed effect of a risk factor on disease status.

Effect Modifier

For example, the effect of smoking on lung cancer was only observed among men (no effect of smoking on lung cancer among women), then gender is a (an) ________________.

Effect Modifier

Which of the following is NOT an appropriate way to increase compliance?

Make a protocol more complicated

All of the following are important criteria when making causal inference except:

Predictive Value

n a clinical trial, researchers might influence the selection of participants into a group (intervention or control), hence introducing bias into the study. What can be done to prevent from this threat to the validity of the study?

Randomization

Suppose a value of 130 mg/dl or higher was considered positive and the prevalence of undetected diabetes remained the same as Q2-Q3. How would you predict Specificity would be affected by this change ? Why?

Specificity would be decreased because FP will increase.

A major difference between an observational study and an experimental study is that something is being done to the participants in an experimental study.

TRUE

Which of the following is the primary aim for establishing inclusion and exclusion criteria in a clinical trial?

To Obtain Valid results

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

Planned cross-over design allows _______ period to reduce carry-over effect from the first assigned treatment.

Wash-out

When should serious consideration be given to using a placebo in a clinical trial?

When the outcome measure is self-perceived pain

Match the study strengths in the left column with study designs in the right column. ____ This is an efficient design that allows us to test two or more hypotheses for the price of one. A. Factorial design ____ This design is useful for reducing bias due to loss to follow-up. B. Run-in design ____ This combines randomized and time-series designs to improve control over confounding, and fewer participants are required to obtain a given level of power. C. Crossover design

A-B-C

Match the descriptions in the left column with the concepts in the right column. ___ Eliminates conscious bias due to physician or patient selection A. Randomize ___ Often impossible in nondrug studies B. Blind ___ More subjective outcomes call for more serious consideration of this C. Placebo ___ Can improve compliance and follow-up ___ Interferes with the doctor-patient relationship ___ Effective at controlling for unmeasured confounding effects

A-B-C-C-A-A

Match the descriptions with the types of screening test results. A. Correctly identifies people as having the disease when, in fact, they do. B. May cause unnecessary stress, anxiety, and treatment. C. May cause a false sense of security. D. The preferred situation to be in.

A. True Positive (TP) B. False positve (FP) C. False negative (FN) D. True negative (TN)

In the study of lung cancer and smoking, air pollution (or any other factor that can cause cancer) can be considered a (an) ___________ variable. A (An) _________________ variable is a variable (say, pollution) that can cause the disease under study (cancer) and is also associated with the exposure of interest (smoking).

Confounding

Which of the following criteria cannot be established using a large, randomized, double-blind experimental study?

Consistency of association

Two similar drugs (drug A and drug B) to alleviate back pain were released simultaneously. An investigator decides to test which of the drugs works better. He decides to conduct a trial in which randomly, half of the participants receive drug A for a month while the other half received drug B. After 2 weeks without any medication, participants switch the drug they were receiving (those who received A start taking B, and those who received B start taking A). What type of study design was illustrated in this example?

Crossover Randomized Clinical Trial

The best index for concluding that an early detection program for breast cancer truly improves the natural history of disease would be:

Improved cause-specific mortality for breast cancer

Based on your literature review, cholesterol level is associated with obesity (exposure). It is also associated with coronary heart disease (outcome). Cholesterol level is:

Potential Confounder

DEFINE ODDS RATIO

Proper measure of association used in case-control study.

Which of the following study designs provides the strongest/weakest support for causal inference?

Randomized controlled trial/ Case report

What is the primary characteristics of an experimental analytic epidemiology study compared with an observational analytical epidemiology study?

Researchers decide the level of exposure among participants

What is the primary distinction between an observational and an experimental analytic epidemiology study?

Researches decide the level of exposure among participants

What are the Se and Sp of the screening test? (Report the answer as a percentage with one decimal place)

Sensitivity = 22.7 Specificity = 99.8

Suppose a value of 130 mg/dl or higher was considered positive, and the prevalence of undetected diabetes remained the same. Then the 2x2 table would have appeared as follows:

Sensitivity = 44.0 % Specificity = 99.0 %

The Tuskegee syphilis study was one reason why the Institutional Review Board was established.

TRUE

A trial is likely to be underpowered, biased, or misleading if which of the following occurs? (Select all that apply)

- A large number of participants are lost to follow-up - A substantial number of participants do not adhere to the study intervention - Baseline variables are not balanced between intervention groups

If a causal association exists, we would expect which of the following?

- An association between an exposure and disease outcome supported in terms of basic human biology. - A biologic gradient not explained by confounding. - An increasing gradient of risk associated with greater risk.

In a case-control study, a researcher found no association between exposure to A and disease B. Later she found out that there was a misclassification when measuring exposure to A, in both cases and controls. What is a potential explanation for the results (no association) of this case-control study?

- Nondifferential misclassification. - The fact that the misclassification affected both cases and controls indicates a nondifferential misclassification, which dilutes the association between A and B.

Smoking participants of a cohort study are more likely to drop out. It is known that the effect of asbestos on getting cancer is much stronger among smokers. When analyzing data for this study, researchers find no association between asbestos and cancer. A potential explanation for the results of the study is

- Selection Bias - Because those participants who dropped the study were particular regarding the association (smokers), those who remained in the study showed a biased attenuated association.

Which of the following improves the reliability of diabetes screening tests?

- Taking more than one sample for each subject and averaging the results. - Having the same lab analyze all samples. - Insuring that the instrument is standardized before each sample is analyzed.

Two neurologists, Drs. J and K, independently examined 70 magnetic resonance images (MRIs) for evidence of brain tumors. As shown in the table below, the neurologists read each MRI as either "positive" or "negative" for brain tumors.

62.9 %

A randomized, double-blind clinical trial of a varicella vaccine observed an estimated incidence of 25% chickenpox episodes in persons receiving the vaccine, compared to 80% among persons receiving a placebo. The estimated efficacy of the vaccine is:

68.8%

What is the efficacy of varciella vaccine?

70-30 / 70 = 57%

Which of the following is an effective way to minimize the threat of observation bias (information bias)?

1. Proper training and blinding of interviewers as to case-control status 2. Use medical records or other recorded health information, rather than recalled information about exposure. 3. Those from which information is being obtained should be unaware of the specific hypotheses being tested.

A group of 10,000 people who were asymptomatic for diabetes were screened for diabetes by a test measuring blood serum sugar content. A value of 180 mg/dl or higher was considered positive. The results were compared with those of a second and definitive test (gold standard) which correctly detected all cases and non-cases. The results were as follows: What is the prevalence of undetected diabetes? (i.e., before screening).

1.5%

This table shows incidence rates (per 100,000) of groups exposed to neither risk factors or to one or two risk factors for lung cancer. What is the expected value of incidence rate X on asbestos exposure group among smokers in additive scale? Incidence Rates of Lung Cancer by Asbestos and Smoking Exposure Status

8.0

Human papillomavirus (HPV) infections are known to be a cause of cervical cancer. In most cases, HPV goes away on its own and does not cause any health problems. However, the worldwide HPV prevalence in cervical carcinomas is 99.7%. This indicates that HPV-negative carcinoma is extremely uncommon. In this case, HPV is which of the following?

A necessary but not sufficient cause

The epidemiologists in charge of the investigation suspects that men in the high exposure work area might tend to overestimate their symptoms (for example, maybe it had been 4 months since they had tingling, or the tingling was of very short duration). She decides to conduct a more objective measure by using an instrument that measures neurological abnormalities. The result of this test indicates that 30% of the exposed workers who had self-reported symptoms (and had been considered diseases) were normal.

Cell a: 42 Cell b: 118 New measures of association: 2.6

Further examination of the test results also reveals that 6% of the unexposed who had no symptoms (initially considered non-diseased), showed abnormalities when tested.

Cell c: 61 Cell d: 329 Measures of assocation: 1.7

Temporal relationship is a necessary criterion to establish causality. Which of the following epidemiologic study design allows to evaluate temporal relationship?

Cohort

Cardiorespiratory fitness, measured as VO2 max in mL/kg/min (maximum oxygen consumption), is associated with increased longevity. Recent evidence suggests that higher fitness increases the life span. Those with a low VO2 max have a life expectancy after 60 years of 7 years, those with moderate VO2 max of 15 years, and those with a high VO2 max have a life expectancy after 60 of 23 years. Which of the causal criteria is illustrated with this example?

Dose-response effect

For an etiologic factor to be judged to be causally related to a disease, which of the following must be true?

Exposure to the etiologic factor must precede the onset of the disease.

In a study that evaluated the association between smoking and colon cancer, the measure of association was OR=3.4 (P < 0.05). You concluded that the association due to chance was not likely. If you were to calculate the corresponding 95% confidence interval, it will include the null value of "1" (one).

FALSE

Epidemiologists tried to investigate the role of factor Y in the exposure-disease relationship. They developed three tables to show the exposure-disease relationship based on the status of factor Y. Based on the information given, what is the role of factor Y in the exposure-disease relationship?

Factor Y is an effect measure modifier

Define cross over clinical trial.

In a crossover clinical trial, participants receive one drug and later switch to another one. The effect of these drugs is compared within the same population.

Investigators think that income might be a confounder in the relationship between Hispanic ethnicity and cardiovascular disease. They conduct a case-control study and they match cases and controls based on income. They find that Hispanics are more likely to have cardiovascular disease as compared with non-Hispanics. What would you conclude about income as a potential confounder in this association? Group of answer choices

Income is not a cofounder

n a case-control study of maternal cigarette smoking as a risk factor for low birth weight, the investigators concluded that mothers of children with low birth weight were more likely to report smoking during pregnancy relative to mothers of children with normal birth weight. The reporting error most likely caused the odds ratio to:

Increase above the true value:

Suppose a value of 130 mg/dl or higher was considered positive and the prevalence of undetected diabetes remained the same as Q2-Q3. How would you predict sensitivity (Se) would be affected by this change? Why?

Increased False Negative (FN)

n a clinical trial, sometimes patients agree to receive an intervention but later decide to stop taking the medication. Moreover, some patients who were assigned to placebo decide to stop taking the placebo pill and buy the drug on their own (even when they do not know that they are taking a placebo pill). When this happens, researchers typically do the analysis according to the original assignment of the treatment regardless of the actual treatment received. What is the name of this procedure to analyze the data?

Intention-to-treat-analysis

Kappa statistics was calculated to measure what extent to which the observed agreement exceeds the agreement expected by chance. The calculation shows that kappa = 0.6. This value of kappa represents what level of agreement?

Intermediate to good

Participants in a clinical trial might report different symptoms when they know they are part of the intervention group instead of the control group. To avoid that, researchers use the following:

Masking (Blinding)

Assume a study investigated the effect of exercise on myocardial infarction, with gender suspected as a confounder. Which method for control of confounding is being used if: A case-control study includes men and women. Controls are selected so that the proportions of male and female subjects between two groups (cases and controls) are identical. Group of answer choices

Matching

In a large case-control study of patients with pancreatic cancer, 17% of the patients were found to be diabetic at the time of diagnosis, compared to 4 % of a well-matched control group (matched by age, sex, ethnic group, and several other characteristics) that was examined for diabetes at the same time as the cases were diagnosed. It was concluded that the diabetes played a causal role in the pancreatic cancer. This conclusion:

May be incorrect because of failure to establish the time sequence between onset of the diabetes and diagnosis of pancreatic cancer (temporality).

Which of the following is a consequence of NOT using a random assignment in a clinical trial?

May overestimate the effect of treatment

Investigators think that income might be a confounder in the relationship between Hispanic ethnicity and cardiovascular disease. They conduct a case-control study, and they match cases and controls based on income. They find that Hispanics are more likely to have cardiovascular disease as compared with non-Hispanics. What is the measure of association used in such study?

Odds Ratio

A retrospective cohort study is carried out to investigate the association between occupational arsenic inhalation and neurological exposure and neurological effects among workers in a copper smelter. For the sake of simplicity, let's assume there are two possible exposure categories: high and low (for example, those working in the smelting process and those working in administration). The outcome was based on self-reported information from an interview that asked: "Have you had tingling in your fingers in the last month that lasted more than 30 minutes?" Those that said "yes" were classified as "diseased", and those that said "no" were the "non-diseased" group. The exposure was carefully assessed by review of company records which reflected very good exposure monitoring (both air sampling and urine testing). In order to avoid selection bias, the company encouraged everyone to participate by telling their workers that they were a concerned employer and wanted to know if there were adverse neurological effects from the potential arsenic exposure in some of the work areas.

RR = 3.7

Suppose the prevalence of undetected diabetes was 2.5% rather than 1.5%. If the same screening test (in the previous question) were used with a value of 130 mg% or higher considered positive, what effect would this increase in prevalence have on:

Sensitivity = No change Specificity = No change PPV = Increased

Assume a study investigated the effect of exercise on myocardial infarction, with gender suspected as a confounder. Which method for control of confounding is being used if: A case-control study includes men and women. The odds ratios are calculated separately for men and women.

Stratification

In a study that evaluated the association between smoking and colon cancer, the measure of association was OR=3.4 and the corresponding p < 0.05. You concluded that the association due to chance was not likely.

TRUE

In an experimental study, double blinding is utilized when accessing outcomes of treatment. The reason for double blinding to reduce bias.

TRUE

In an experimental study, participants are randomly allocated (by chance, rather than by choice), to two different experimental groups. The reason for randonmization is to reduce bias.

TRUE

Two main types of systematic errors are selection and information bias.

TRUE

In a hypothetical cross-sectional study, marriage was found to be strongly associated with increased levels of happiness. Authors concluded that marriage was a cause of happiness. However, a reviewer of this study argued that because it is a cross-sectional study, it could not be determined whether marriage caused happiness or happiness caused marriage. What criterion of causal relationships was not met in the study according to the reviewer?

Temporal relationship

What is the purpose of randomization?

The main purpose of randomization is to prevent the potential bias on the part of the investigators from assigning participants into the groups of the trial.

In a published epidemiologic study investigating infertility related to sexually transmitted diseases (STD), the authors state that 5% of identified cases refused enrollment, 10% of identified cases were lost to follow-up prior to data collection, and 10% of interviewed cases had missing data for one or more key variables describing exposure. Based on this information, which of the following statements is most likely to be true?

There is a potential for selection bias which could be differential with respect to cases and controls

Compare the results of your calculations from each of the 3 tables above (Q8, Q9, & Q10). In what direction was the bias operating? Hint: Which one (Q8, Q9, or Q10) is true association (without misclassification bias) ? and Which one (Q8, Q9, or Q10) is the biased results (with misclassification bias) ? Compare truth to biased results.

Toward the conclusion that the exposure was associated with a greater risk of the disease outcome (away from the null)