Prev Med Final Block 2

Selection of an inappropriate control group would be which type of bias? a. control selection bias b. self-selection bias c. differential referral, surveillance, or diagnosis bias d. loss to follow-up

A. Need to derive controls exposed to the same variables as the cases

This type of internal validity answers the question: does the measure as a whole test what it says it's going to test? a. content validity b. criterion validity c. construct validity

A; Content validity

This term describes the relative absence of bias or systematic error. a. validity b. reliability c. precision

A; aka accuracy

You are performing a study on the prevalence of substance abuse among a population. You use date "anchors" such as holidays to gauge when their last use of a certain drug was. This type of questioning is to reduce: a. recall bias b. interview bias c. procedure bias d. misclassification bias e. observer-expectancy bias

A; recall bias

what Parametric tests compares the means between three or more independent groups? ex: Test the difference in total cholesterol level among patients who took Lipitor, Crestor or generic drug

ANOVA. Parametric so normally distributed ANalysis Of VAriance

what is the difference between Validity and Reliability?

Accuracy (Validity): How close the measurement is to the real Value Precision (Reliability): Consistency and Reproducibility of the measurement

If all studies with a given relation produce similar quantitative results, causal inference is enhanced. This applies to which aspect of Bradford Hill's Causal Guidelines? a. strength of association b. dose-response gradient c. temporality d. consistency of findings e. biological plausibility

D

This aspect of Hill's Causal Guidelines answers the question: "Is there another scenario/disease that we can relate this to?" a. strength of association b. dose-response gradient c. specificity of association d. analogy e. biological plausibility

D

This type of information bias is most common in epidemiology research and erroneous classification of exposure or disease status. a. recall bias b. interview bias c. procedure bias d. misclassification bias e. observer-expectancy bias

D

People are genetically different from each other. Because of this, which type of data variation could be the reason for error?

biological variability

what term describes the degree to which the results of an observation hold true in other settings?

External Validity aka Generalizability

In a study of low back pain, researchers reviewed the medical records of 40 men with low back pain under care at their clinic - 22 had been referred for surgery and the other 18 were under care without surgery. The study compared rates of disabling pain after 2 months. Rates of pain were slightly higher in the surgically treated patients. Which threat to validity is present, Internal Validity or External Validity?

External Validity due to generalizability

True or false? Selection bias does not occur in retrospective cohort studies.

False; selection bias can occur in any type of experiment

what kind of bias is the Healthy Worker Effect an example of?

Healthy worker effect is a form of self-selection bias. There's a self-screening process, people who are unhealthy will screen themselves out of active worker population.

When studying for the effects/presence of bias, what must investigators do?

Identify the source of bias, Estimate its magnitude or strength, and Assess its direction

In a direction of bias graph, what are the two options that the association of bias could be?

If the bias is on the right side of the null (line of no effect), that is a positive association which could either be toward or away from the null depending of the true value. The image shows a Positive Association toward the null value. If the bias and true value are on the left of the line of no effect, that is a Preventative association. This could also be either toward or away the null value with respect to the true value.

what is Control Selection Bias and what kind of study is this seen in?

In a case-control study selection bias occurs when subjects for the "control" group are not truly representative of the population that produced the cases. Ex: Cases were derived from the hospital and controls came from the neighborhood around the hospital

How would you calculate Relative Risk if given a 2x2 table?

RR = (a/(a+b))/(c(c+d)) = EER/CER o Note: Here, the "exposed" (numerator) is called the experiment event rate (EER), and the "unexposed" (denominator) is called control event rate (CER)

In general, how do you calculate Relative Risk?

RR= Incidence of outcome in the exposed / Incidence of outcome in unexposed RR=1 (no association) RR>1 (risk in exposed greater than risk in non-exposed) RR<1 (risk in exposed less than risk in non-exposed)

what is the difference between Systemic and Random error?

Random error (non-differential): use of invalid outcome measure that equally misclassifies cases and controls. Systematic (differential error): use of an invalid measures that misclassifies cases in one direction and misclassifies controls in another

In which type of study is it most common to see loss to follow-up?

Randomized Control Trial

what is the definition of Reliability? (lack of ______)

Reliability: defined as relative lack of random error. Validity is defined as relative absence of bias or systematic error.

what does Observer-Expectancy Bias describe?

Researcher's belief in the efficacy of a treatment changes the outcome of that treatment

An epidemiologist examines the association between the use of diet pills and migraine headaches using a case-control study design. His subjects are outpatients at a large community hospital. He finds that those using diet pills are more likely to complain of migraine headaches but he also discovers that those who have migraines headaches and use diet pills are more likely to be referred to the community hospital for outpatient diagnostic testing than those who have migraines but are not using diet pills. What type of bias is present in this study?

Selection Bias

In a study of low back pain, researchers reviewed the medical records of 40 men with low back pain under care at their clinic - 22 had been referred for surgery and the other 18 were under care without surgery. The study compared rates of disabling pain after 2 months. Rates of pain were slightly higher in the surgically treated patients. The patients who were referred for surgery were younger and more fit than those who remained under medical care. What type of bias is present?

Selection Bias

A researcher was conducting a case control study on levels of stress and cardiovascular disease. Interviews could only be scheduled from 10:00am to 2:00pm; working women were less likely to participate because they were not available to be interviewed during those time periods. What kind of bias occurred?

Selection bias

what are the two main types of Systemic bias?

Selection bias and Information bias

what kind of bias is introduced when the study population does not represent the target population? what two kinds of studies is this bias more common in?

Selection bias is introduced when the study population does not represent the target population. Selection bias could be more common in a retrospective cohort study and case control

why is selection bias more likely to occur in Case-Controls and Retrospective Cohorts?

Selection bias is more likely to occur in case-control and retrospective cohort studies because both the exposure and disease have occurred by the time of subject selection

what is the main way to reduce Misclassification Bias?

Since this is a measurement error bias, the main way to reduce is to improve the accuracy of the collected data. This can by done by validation of data or by using more sensitive/specific criteria

In the analysis phase of a study, how would you control for confounding variables?

Stratification (subgrouping), Standardization (equalize weight given to variables), and Adjustments (multivariate analysis)

What methods were used to control confounding bias during data analysis in the SDL OMM research paper?

Stratification by co-treatment and analyzed secondary outcomes to observe trends Compared both intention to treat analysis to per protocol analysis Examined non-responder data as well

What method is a control for confounding primarily used in the analysis stage?

Stratification. Stratification is the only technique that reduces confounding during the analysis stage. It involves breaking the data into strata that can be more descriptive.

what are the 6 types of Causal relationships?

Sufficient Cause: If the factor (cause) is present, the effect (disease) will always occur. Necessary Cause: The factor (cause) must be present for the disease to occur; however, a necessary cause may be present without the disease occurring. Risk Factor: If the factor is present, the probability that the effect will occur is increased. Direct causal association: The factor exerts its effect in the absence of intermediary factors (intervening variables). Indirect causal association: The factor exerts its effect through intermediary factors. Non-causal association: The relationship between two variables is statistically significant, but no causal relationship exists because the temporal relationship is incorrect

what is the definition of Effect Modification?

The degree in which the association between exposure and outcome varies between groups. motivation behind the assessment of effect modification is to identify whether the effects of a treatment (or exposure) is different in different groups with different characteristics.

When looking at a Forrest Plot, what in there tells you the size of effect?

the confidence intervals

If you are wanting to compare the means between two independent samples that are normally distributed, what test would you use?

unpaired t-test

Which of the following is not one of the Bradford Hill criteria for causation? A. Dose-response B. Statistical significance C. Reversibility D. Biologic plausibility E. Analogy

C

A paper on lead exposure indicates that toxic effects on the central nervous system of infants and children are known, however it is unknown if these effects persist into adulthood. This study extended its follow-up to ages 28 to 30 years. A previous assessment of the same population found that early childhood blood lead levels are associated with cognitive function in adolescence. If this information is in a publication, what aspect/part of the study are the authors communicating? A. Introduction B. Study Objective C. Study Context D. Analysis E. Results

C. Study Context

How would researchers in a study that has Confounding prevent it next time? a. Chi-square test b. Control group that better represents the general population c. Better matching among study groups. d. Decreasing the study power e. Increasing the sample size

C. better matching among study groups

This term determines the reproducibility of an experiment. a. accuracy b. validity c. precision

C; aka reliability

Randomization and blinding can be done to reduce: a. recall bias b. interview bias c. procedure bias d. misclassification bias e. observer-expectancy bias

C; procedure bias

What group helps alleviate the problems arising from inadequate reporting of randomized controlled trials (RCTs)?

CONSORT, which stands for Consolidated Standards of Reporting Trials

A study examining 450 workers in the cotton textile industry to examine asbestos and lung cancer. There were 222 workers exposed to asbestosis and of those 142 have lung cancer and 80 do not. 228 workers were not exposed to asbestosis and 74 of those have lung cancer and 154 do not. What is an appropriate statistical analysis to use to test for an association between asbestosis exposure and lung cancer?

Chi-Square Test

A cohort study was conducted to examine the association between tobacco smoking and lip cancer. Women who reported wearing lip gloss more frequently were less likely to report a history of tobacco smoking. In a previous study, wearing lip gloss was also associated as protective factor against lip cancer. Further tobacco smoking increases the risk of lip cancer. what type of bias/error is this?

Confounding

A researcher is studying the effects of esophageal adenocarcinoma in smokers. The results show a positive correlation between smoking and esophageal adenocarcinoma. The researcher then removed all of the alcoholics from the group under study. The new results showed no correlation between smoking and esophageal adenocarcinoma. Which type of bias most likely exists in this second study?

Confounding

If experimental values exhibit a relative risk (RR) of 0.5, but the actual relative risk of the experiment should have been 0.3, what kind of bias is this? a. overestimation b. underestimation

A, Overestimation

This type of causal relationship states: if the cause is present, the disease will always occur. a. sufficient cause b. necessary cause c. risk factor d. direct causal association e. non-causal association

A

Very increased differences in relative risk may be a strong argument for causation, which can be described by which aspect of Bradford Hill's Causal Guidelines? a. strength of association b. dose-response gradient c. temporality d. consistency of findings e. biological plausibility

A

Which aspect(s) of Bradford Hill's Causal Guidelines may be examples of confounding variables? a. strength of association b. dose-response gradient c. temporality d. consistency of findings e. biological plausibility

A and B

A chi-squared test is used to determine if a significant difference exists in two populations where both the dependent and independent variables are __________________

A chi-squared test is used to determine if a significant difference exists in two populations where both the dependent and independent variables are categorical

How is a confounder different from effect modification?

A confounder is something that gets in the way, while effect modification is a stratification (dividing) into subgroups to see how data differs between those groups

This type of selection bias can occur in a case-control study as a result of differential tendency to oversample cases based on their exposure. a. control selection bias b. self-selection bias c. differential referral, surveillance, or diagnosis bias d. loss to follow-up

C

"Disease frequency increases or decreases monotonically with the level of exposure" is described by which aspect of Bradford Hill's Causal Guidelines? a. strength of association b. dose-response gradient c. temporality d. consistency of findings e. biological plausibility

B

The Healthy Worker Effect is which type of bias? a. control selection bias b. self-selection bias c. differential referral, surveillance, or diagnosis bias d. loss to follow-up

B

This type of causal relationship states: the cause must be present for the disease to occur, although it does not always result in disease. a. sufficient cause b. necessary cause c. risk factor d. direct causal association e. non-causal association

B

If experimental values exhibit a relative risk (RR) of 1.4, but the actual relative risk of the experiment should have been 1.9, what kind of bias is this? a. overestimation b. underestimation

B, underestimation

How would researchers in a study that has Selection Bias prevent this next time? a. Chi-square test b. Control group that better represents the general population c. Better matching among study groups. d. Decreasing the study power e. Increasing the sample size

B. a control group that better represents the general population can prevent selection bias, but does not affect statistical power of a study.

Which of the following choices describes the causal factor occurring before the disease developed in a study? a. elimination b. temporal association c. causation association

B; Temporal association

This type of internal validity answers the question: do scores measure the variables being measured? a. content validity b. criterion validity c. construct validity

B; criterion validity

You are asking a patient a question, but she does not understand what you are asking her. You then begin to list different answers she can provide to answer the question you asked. This can lead to which type of bias? a. recall bias b. interview bias c. procedure bias d. misclassification bias e. observer-expectancy bias

B; interview bias

what three alternative explanations must be eliminated before a study is considered valid?

Bias, Confounding, and Random Error must be eliminated for a study to be considered valid. These are also the 3 major parts of internal validity.

This is the most prominent model of causation.

Bradford Hill's Causal Guidelines

"Exposure must precede the disease" is described by which aspect of Bradford Hill's Causal Guidelines? This can be argued by parents who say vaccines cause autism--but were the signs of autism present before the vaccinations were given? a. strength of association b. dose-response gradient c. temporality d. consistency of findings e. biological plausibility

C

The more narrow/specific the association is, the stronger the causal relationship may be. For example, smoking, asbestos exposure, and radiation exposure are related to lung cancer. Which aspect of Bradford Hill's Causal Guidelines can this be applied to? a. strength of association b. dose-response gradient c. specificity of association d. consistency of findings e. biological plausibility

C

Which of the following is NOT an essential attribution to causation? a. association b. time (must precede event) c. direction d. all of the above are essential to causation

D. Association, Time, Direction are all essential to causation

A recent cohort study was conducted to observe the relationship between vitamin E megadosing and myocardial infarction. The researchers conducted a correlation analysis and reported a correlation coefficient (r) of -0.95. What is the appropriate conclusion? A. A correlation cannot be determined using a cohort study B. vitamin E megadosing correlation with incidence of myocardial infarction cannot be determined with a correlation coefficient less than 1 C. vitamin E megadosing has no correlation to incidence of myocardial infarction D. vitamin E megadosing is correlated to decreased risk of myocardial infarction E. vitamin E megadosing is correlated to increased risk of myocardial infarction

D. A negative correlation, such as -0.95, would indicate that vitamin E megadosing is correlated to decreasing incidence of myocardial infarction

what are different ways to reduce Recall Bias?

Design structured questionnaires, use self-administered questionnaires, used biological measurements and pre-existing records, and mask participants to study hypothesis.

It has been suggested that physicians may examine women who abuse opioids more thoroughly than women who do not. If opioid abuse is found to be associated with an increase in STD rates, what type of bias would be present?

Differential Selection Bias

what is Differential Surveillance?

Differential Surveillance is a type of selection bias can occur in a case-control study as a result of differential tendency to oversample cases based on their exposure.

This aspect of Bradford Hill's Causal Guidelines states that scientific or biological/chemical knowledge may have a causal interpretation. a. strength of association b. dose-response gradient c. temporality d. consistency of findings e. biological plausibility

E

A study is conducted to determine the mean cholesterol levels between smokers and non-smokers. After the data is collected, it is determined that they are continuous variables and normally distributed. The researchers want to test the difference in the means of the two study groups. Which of the following is the most appropriate test to analyze the study? A. analysis of variance B. chi-squared C. linear regression analysis D. standard deviation analysis E. t-test

E. t-test

how are Effect Modification and Confounding different?

Effect modification is distinct from confounding; it occurs when the magnitude of the effect of the primary exposure on an outcome (i.e., the association) differs depending on the level of a third variable.

In an example: Smoking is associated with pancreatic cancer Smoking is associated with coffee drinking Smoking is not a result of coffee drinking What is the term to describe smoking in this scenario?

In this scenario, Smoking is a confounding variable.

An epidemiologist performed a case-control study to determine the risk for cervical cancer associated with a history of sexual promiscuity. The results measured an odds ratio of 0.98 with a p-value of 0.06. It was later revealed that a type II error occurred. How could this error have been prevented?

Increasing Sample Size. Type II error (false negative) occurs when there is low statistical power. One way to increase sample size could be a meta-analysis and combine data from many studies.

what is an independent and dependent variable?

Independent: the variable that represents the quantity manipulated in an experiment Dependent: the quantity whose value depends on how the independent variable is manipulated

what is the difference between Information Bias and Selection Bias?

Information bias is an error that arises from systematic differences in the way information on exposure and disease is obtained from study groups. Results in participants who are incorrectly classified as either exposed or unexposed. A selection bias is an error that arises from systematic differences in selecting participants into the study and following the study groups through time

what term describes the methodologic rigor of a trial, which is dependent on the elements of the trial design?

Internal Validity

what test would you used to test the differences in continuous variable in more than 2 groups when data is not normally distributed?

Kruskal-Wallis test the non-parametric alternative to ANOVA

What provisions were made to minimize the influence of confounding factors prior to the analysis of the data in the SDL OMM research paper?

Measured baseline characteristics Randomization Documented other low back pain care patients were receiving Specific inclusion and exclusion criteria

what kind of study is Loss to Follow Up a major concern in?

Prospective Cohort studies are very concerned about Loss to Follow Up. Over 80% follow-up rates are the goal

what was the study design of the SDL OMM research paper?

RCT, double-blind, sham-controlled, 2x2 factorial design

what was the study population in the SDL OMM research paper?

Non pregnant patients ages 21-69 years of age who had low back pain for at least 3 months were recruited in Dallas-Fort Worth, Texas. 455 total

what is used from the data to perform statistical analysis in Nonparametric statistics versus Parametric statistics?

Nonparametric statistics: Median is used for analysis Parametric statistics: Use the Mean of the data to perform statistics

what were the author's main conclusions from the SDL OMM research paper?

OMT was effective in providing moderate to substantial improvements in low back pain and decreased the amount of concurrent use of prescription drugs UST was not effective in improving low back pain

How would you calculate Odds Ratio if given a 2x2 table?

OR= a*d / b*c OR = 1 (odds of exposure equal among cases and controls) OR > 1 (odds of exposure higher among cases) OR < 1 (odds of exposure lower among cases)

What measure of association will you calculate in a case-control study?

Odds Ratio is the measure of association used for a case-control study

what is the outcome variable for the unpaired t-test?

Outcome variable for unpaired t-test is continuous; the independent variable is categorical

If you are wanting to compare the means between two dependent samples that are normally distributed, what test would you use?

Paired t-test

A study is comparing the correlation between two continuous variables that are normally distributed. What type of test would you use to assess the magnitude and direction of the relationship?

Pearson

After ANOVA testing has been done on 3 or more independent groups, what additional testing can be done to ascertain where the differences are found?

Post Hoc test

Name the 5 types of Information Bias.

Procedure Bias Recall Bias Interviewer Bias Misclassification Bias Observer-expectancy bias PRIMO

what types of error should the term "bias" be reserved for?

The term 'bias' should be reserved for differential/systematic error. Random error (non-differential): use of invalid outcome measure that equally misclassifies cases are controls. Systematic (differential error): use of an invalid measures that misclassifies cases in one direction and misclassifies controls in another

what is Interviewer Bias?

Type of Information bias where there is systematic difference in soliciting, recording, interpreting information

what is Recall Bias?

Type of information bias where people who have experienced an adverse event (disease) are more likely to recall previous risk factors than people who never experienced the event

what is Procedure bias?

Type of information bias where subjects in different groups are not treated the same

what is Misclassification bias? what type of study is this most common in?

Type of information bias where the subject's exposure or disease status is erroneously classified. This is the most common type of bias in epidemiology research.

An obesity researcher collected data for 3 months on fast food consumption by gender (male and female). Distribution of the dependent variable (serum cholesterol) was normally distributed. What is an appropriate statistical analysis for this data?

Unpaired t-test

what is the definition of Validity? the lack of what?

Validity is defined as relative absence of bias or systematic error. Reliability is the relative lack of random error.

A study is comparing the medians, and is comparing two dependent samples. What type of test would you use to assess the differences between the groups?

Wilcoxon

A researcher used the medians of his data to compare shoulder function before and after surgery orthopedic surgery on the shoulder. What type of analysis did the researcher most likely used?

Wilcoxon Signed Rank Test

If you are wanting to compare the median between two independent samples that are NOT normally distributed, what test would you use?

Wilcoxon rank sum test

If you are wanting to compare the median between two dependent samples that are NOT normally distributed, what test would you use?

Wilcoxon signed rank test

A study is conducted to determine the mean cholesterol levels between smokers and non-smokers. After the data is collected, it is determined that they are continuous variables and normally distributed. The researchers want to test the difference in the means of the two study groups. Which of the following is the most appropriate test to analyze the study?

t-test

Does a larger sample size negate bias?

no

what are the two tests used to estimate the degree of association between two quantitative variables?

non-parametric) Spearman & (parametric) Pearson Shows direction and magnitude of relationship (correlation)

what is a "web of causation?

schematic of how agents work together to cause disease (particularly used in chronic diseases/noninfectious diseases)

what was the purpose of the SDL OMM research paper?

studied the efficacy of osteopathic manual treatment (OMT) and ultrasound therapy (UST) for chronic low back pain

What are the three possible reasons for variation in data? Which is the most common?

• Biological variability error - People are different • Imprecision/Experimental error • Mistakes and glitches - Simple mistake (something written down wrong, etc.) Experimental error is considered the primary reason for variability in statistics, but in science, biological variability is considered the main contributing factor.

What are the three criteria that must be proven in critical analysis?

• Effectiveness of clinical interventions, accuracy and precision of diagnostic tests, and the power of prognostic markers

What type of evidence increases the conviction in a cause-and-effect relationship?

• Empiric evidence

What type of statistical test uses a contingency table and is used only when there are small samples, having typically a cell count of less than 5 count?

• Fisher's exact test

What specific type of selection bias is of concern in occupational cohort studies?

• Healthy worker effect (which is a type of self-selection bias)

When studying the presence of bias, investigators must do what three things?

• Identify the source, estimate its magnitude/strength, and assess its direction

What is power? How is it determined?

• Power is the test's ability to correctly reject the null hypothesis • It is determined by sample size

If your experiment has a relative lack of random error, what does it have?

• Reliability

What is the difference between selection and information bias?

• Selection bias - An error that arises from systematic differences in selecting participants into the study and following the study groups through time. • Information bias - An error that arises from systematic differences in the way information on exposure and disease is obtained from study groups

what type of study is Self-Selection bias mostly seen in?

• Self-selection bias - Refusal, nonresponse, or agreement to participate that is related to the exposure and disease. This is mostly seen in case-control studies

What are Bradford Hill's Causal Guidelines (BHCG)?

• Strength of association (effect size) - Expressed by a large relative or absolute risk • Dose-response gradient (biologic gradient) - Disease frequency increases or decreases monotonically with the level of exposure • Temporality (if cause, then effect) - Exposure must precede the disease • Consistency of findings (reproducibility) • Biological plausibility of hypothesis (greater exposure should lead to greater effect) • Coherence of evidence (epi + lab) - Findings do not seriously conflict with our understanding of the natural history of the disease • Specificity of association (specific population) - Specific exposure is found to be associated with only one disease • Experimental evidence • Analogy with known effects

What type of error misclassifies cases in one direction and controls in another? What type of error misclassifies cases as controls, leading to imprecision in measures of association?

• Systematic (differential error) misclassifies cases in one direction and controls in another • Random (non-differential error) misclassifies cases as controls.

What is given in the introduction of a study?

• The rationale for the study (an explanation of what the study investigates and why).

What is the generic structure of a journal article?

• Title and Abstract, Introduction, Methods, Results, Discussion, Key results, Limitations, and Interpretations

If your experiment has a relative absence of bias or systematic error, what does it have?

• Validity

What are the different ranges for interpreting correlation?

• Weak Correlation - Values between 0 - 0.3 • Moderate Correlation -Values between 0.31 - 0.7 • Strong Correlation - Values greater than 0.7 Pearsons/Spearmans

What is the difference between an α level and a p-value?

• α level - The highest risk of making a false-positive error that the investigator is willing to accept (usually set to 0.05) • p-value - Gives the probability of obtaining the observed result by chance rather than as a result of a true effect.

What is the difference between a type I and type II error?

•Type 1 Error: Rejecting the H0, when the null is actually true o False positive •Type 2 Error: (Accept) Fail to reject the H0, when the null is actually false o False negative