Biostats/Epidemiology

Question 1

How do you calculate case fatality rate?

Answer 1

CFR = Divide the number of fatal cases of a disease or condition by the total number of people with that disease or condition

Question 2

How do you calculate absolute risk reduction?

Answer 2

Dexmedetomidine (treatment group): 300 cases per 1,000 patients (or 300 / 1,000)

Saline (control group): 500 cases per 1,000 patients (or 500 / 1,000)

These results indicate that the absolute risk reduction (ARR) (ie control rate - treatment rate) in ED between groups is (500 / 1,000) - (300 / 1,000) = 0.20.

ARR = 0.20 indicates that 20% of patients did not develop ED as a result of having received dexmedetomidine rather than saline.

Question 3

How do you calculate positive predictive value (PPV)?

What does it tell you?

What do predictive values depend on?

Answer 3

Positive predictive value is the probability that an individual has a disease given a positive test.

It is calculated as follows: true positives / (true positives + false positives).

Predictive values depend on the prevalence of the disease in the study population; as the disease prevalence increases, PPV increases and NPV decreases, and vice versa.

Question 4

What does the specificity of a test refer to?

How is it calculated?

Answer 4

The specificity of a test is its ability to correctly identify individuals without the disease.

Specificity can be calculated as follows: Specificity = True negatives / (True negatives + False positives)

Question 5

What does prevalence of a disease refer to?

What is the difference between point prevalence and period prevalence?

Answer 5

Prevalence refers to the proportion of diseases individuals in a particular at risk population

Period prevalence refers to the number of disease cases in a period (eg, from July to July 31) divided by the number of people in the at-risk population (ie, prevalent cases at the beginning of a period plus any incident cases during the period).

Point prevalence refers to the number of disease cases that are active at a specific point in time (eg, July 31) divided by the number of people in the at-risk population.

Question 6

What does relative risk (RR) measure?

How is calculated?

What kind of studies report RR?

Answer 6

Relative risk is the measure of association between exposure to a risk factor and an outcome or disease.

Cohort or experimental studies

Question 7

How is number needed to treat (NNT) calculated?

What does a lower NNT indicate?

Answer 7

NNT = 1/ARR

NNT = 1 divided by Absolute Risk Reduction (ARR)

Lower NNT = more effective treatment

Question 8

What does the absolute risk reduction describe?

Answer 8

ARR describes the difference in risk of unfavorable outcome between the treatment group (new treatment) and the control group (standard treatement)

Question 9

How do you calculate cumulative incidence of a disease?

Answer 9

The cumulative incidence of a disease is the number of new cases of a disease over a specific period divided by the total population at risk at the beginning of the study (ie, the proportion of at-risk individuals who contract the disease over the specified period).

Question 10

What are the three types of population pyramids?

Answer 10

There are three types of population pyramids: expansive (ie, young and growing population; high birth and mortality rates), stationary (ie, stable population; declining birth rates and low mortality rates), and constrictive (ie, shrinking population; significantly low birth and mortality rates).

Question 11

What is a case control study?

Answer 11

Observation study design where individuals are selected who have the outcome (cases) and individuals who do not have the outcome (controls) and then retrospectively comparing the history of the exposure to risk factors.

Question 12

How are subjects in a cohort study selected?

Answer 12

1. Subjects are initially selected for the population of interest
2. These subjects are identified as exposed or not exposed (exposure status to risk factor) - identified according to independent variable
3. After catogorized bsed on their exposure status, the occurence of the dependent vriable (outcome of interest) over a specific period is determined in each group

Question 13

When do you use a t-test to compare the mean?

When do you use a ‘Analysis of variance’ (ANOVA) to compare the mean?

Answer 13

A t-test is used to compare the difference between the means of 2 groups.

Analysis of variance (ANOVA) compares the difference between the means of 2 or more groups.

Results from a t-test and ANOVA test will be equivalent when comparing the difference between the means of 2 groups.

Question 14

What are the phases of clinical trials?

Answer 14

Phase I - small, healthy participants, assess safety and pharmokinetics, often preformed in a controlled environment

Phase II - small to medium, participants have the condition of interest, assess treatment efficacy, toxicity, adverse effects, optimal dosing strategies

Phase III - large trials (typically over 300), fully assess treatment response and safety, these trials must show adequate effetiveness compared to standard treatment for teh drug to obtain regulatory approval

Phase IV - after drug has obtained regulatory approval for clinical use, long term beneftis and risks or identify uncommon adverse affects

Question 15

What is RR by definition?

How is it calculated?

Answer 15

RR = risk among the exposed/risk among the unexposed

RR = (outcome present/total exposed)/(outcome present/total unexposed)

RR = [a/(a+b)] / [c/(c+d)] = (120/400)/(100/300) = 0.90

Question 16

When should you consider lead-time bias during evaluation of a study design?

When does lead time bias occur?

Answer 16

• always consider when evaluating any screening test
• Lead-time bias occurs when a new test diagnoses a condition earlier than conventional studies, causing an apparent increase in survival time despite no improvement in overall mortality.
• Long-term mortality rates, not survival times, should be considered for measuring the effect of early screening and treatment.*

Question 17

What is the Hawthorne effect?

Answer 17

The Hawthorne effect (observer effect) is the tendency of study subjects to change their behavior as a result of their awareness that they are being studied.

This can impact the observed outcomes, thereby seriously affecting the validity of the study. The Hawthorne effect is commonly seen in studies concerning behavioral outcomes or outcomes that can be influenced by behavioral changes. In this example, physicians (not patients) are the subjects of the study; those physicians who are aware that they are being studied may modify their behavior and start taking sexual histories. To minimize the Hawthorne effect, study subjects can be kept unaware that they are being studied, but this can occasionally pose ethical problems.

Question 18

What is the concept of accumulation effect?

Answer 18

This concept of accumulation effect can apply to both risk factors and risk reducers. The effect of exposure to risk factors may depend on the duration and intensity of the exposure; long-term exposure may be necessary well before an effect on the disease process is clinically evident (eg, lung cancer developing after decades of smoking exposure). Similarly, exposure to certain risk reducers must occur continuously over extended periods before disease outcome is affected. In this case, >5 years of continuous antioxidant use (risk reducer) were required to reveal their protective effect on stroke.

Question 19

What is relative risk reduction?

How do you calculate relative risk reduction (RRR)?

Answer 19

The percent reduction in absolute risk (AR) between treatment and the control group (standard therapy).

Relative risk reduction =

(absolute risk_control - absolute risk_treatment) / absolute risk_control

RRR may overstate the effectiveness of an intervention. For example, a RRR of 50% occurs whether a drug decreases the incidence of a disease from 2% to 1% or from 50% to 25%. Clearly, the latter is of greater clinical significance.

Question 20

When do you use analysis of variance (ANOVA) to compare values vs a Chi square test?

Answer 20

ANOVA - compares means of 3 or more groups >> quantitative

Chi-square test evaluates teh associate between 2 categorical values >>> qualitative

Question 21

What are you comparing in a case control study?

What are appropriate measures for comparison?

Answer 21

Once cases and controls are identified, the frequency of past exposure to >1 risk factors of interest is compared between cases and controls to estimate the association between the risk factors and the outcomes.

Therefore, an appropriate measure for the proposed study would be any event that preceded HCV infection. Among the given choices, a history of past blood transfusions precedes HCV infection.

Case vs control identified according to their disease status

Question 22

What measure of association would a cohort study most likely report?

Answer 22

The relative risk = risk of disease in exposed/risk of disease in nonexposed

EXPOSURE STATUS TO A RISK MODIFIER (vaccinated vs unvaccinated)

Question 23

What does relative risk reduction measure?

What are the methods to calculate it?

Answer 23

RRR measures how much a treatment reduces the risk of an unfavorable outcome

RRR = (Risk_control - Risk_treatment) / Risk_control

RRR = 1 - relative risk

RR = ( Risk_treatment / Risk_control)

Question 24

What does a normal distribution curve (Gaussian) tell you about what percent of observations lie within 1/2/3 standard deviations of the mean?

Answer 24

Question 25

What measurements are sensitive/not sensitive to outliers in a dataset?

Answer 25

An outlier is defined as an extreme and unusual observed value in a dataset. It can affect measures of central tendency (mean, median, mode) as well as measures of dispersion (standard deviation, variance).

>>> Modes tend to be resistant to outliers (mode is the most frequently observed data point)

Question 26

How are mean, median and mode shifted in a positively skewed deviation and a negatively skewed deviation?

Answer 26

• positively skewed distribution, the mean is the most shifted in the positive direction, followed by the median and then the mode (mode < median < mean), the median often reflects a central tendency better than the mean does in this situation
• negatively skewed distribution (with the “tail” on the left), the mean is the most shifted in the negative direction, followed by the median and then the mode

Question 27

How do postive and negative predicative values depend on disease prevalence in a tested population? Sensitivity and specificity?

Answer 27

PPV and NPV depend on disease prevelance

PPV is direclty correlated

NPV inversely correlated

Sensitivity and specifcity are not afffected

Question 28

How does the degree of overlap between healthy and diseased population curves affect sensitivity and specificity of a quantitative diagnostic test?

Answer 28

The degree of overlap between the healthy and the diseased population curves limits the maximum combined sensitivity and specificity of a quantitative diagnostic test. The degree to which sensitivity or specificity is affected depends on the chosen cutoff value.

The corresponding decrease in the number of FPs and FNs means the new serum marker has higher sensitivity and specificity (ie, better performance) in women with BRA mutations.

Question 29

What does the ‘power’ of a test tell you?

What does the p-value tell you about observed results?

Answer 29

The power of a test is the probability of making the correct decision of rejecting a false Ho (ie, determining there is a correlation when one truly exists).

The p-value is the probability of obtaining the observed result (or results more extreme) when H, is assumed to be true; it is informally interpreted as the probability that the observed results are due to chance.

Question 30

What is the probability that all events will turn out the same if the events are independent events (one pateint’s result has no impact on another’s)?

What is the probability at least 1 event will turn out differently?

Answer 30

If events are independent, the probability that all events will turn out the same is the product of the separate probabilities for each event. The probability of at least 1 event turning out differently is given as 1 P(all events being the same).

Question 31

What type of study uses the odds ratio as a measure of association?

What is the null value of an odds ratio?

Answer 31

The odds ratio (OR) is a measure of association used in case-control studies, cases by definiton already have the disease.

It quantifies the relationship between an exposure and a disease; its null value (ie, null hypothesis value) is always 1 (ie, OR = 1).

** Null hypothesis is a statement of no difference or no association

Question 32

How does the significance level alpha in a study effect the confidence of a study?

Answer 32

Reducing the significance level alpha (a) in a study allows researchers to report any significant findings with greater confidence.

Alpha is also the probability of making type I error, which is rejecting a true Ho (ie, a false positive: finding a statistically significant difference when one does not truly exist).

Question 33

How does a higher or lower cutoff value for a diagnostic test affect sensitivity and specificity?

Answer 33

Highly sensitive test - negative test will rule out the disease (SnNOut)

Highly specific test- a positive result will rule in the disease (SpPIn)

Question 34

What is the multiplication law of probability?

Answer 34

The multiplication law of probability states that the probability of 2 or more independent events occurring together can be calculated by multiplying the individual probabilities of each event. The multiplication law of probability can be extended to complementary events (1 probability of event) in the same manner.

Question 35

When should you consider a highly sensitive test vs highly specific test?

Answer 35

The sensitivity of a test refers to its to ability to correctly identify those with the disease. A highly sensitive test should always be considered over a highly specific test when screening for life-threatening diseases, where identification of every person with the disease is important.

Question 36

How do you caclulate the odds ratio?

Answer 36

OR = (odds of exposure in cases)/(odds of exposure in controls)

OR = (a/c)/(b/d)

Question 37

What are the different levels of health prevention?

Answer 37

Question 38

How is the relative risk interpreted?

Answer 38

Question 39

How will a factor affecting duration of disease affect the prevalance of disease?

Answer 39

Prevalence equals the incidence rate multiplied by the average disease duration. Changing disease prevalence in a steady-state population with a constant incidence rate means that there is an additional factor affecting the duration of the condition. A factor that prolongs disease duration (eg, improved quality of care) will increase disease prevalence, as affected patients survive longer.

Question 40

What is the attack rate?

Answer 40

The attack rate is the ratio of the number of people who contract an illness divided by the number of people who are at risk of contracting that illness.

Question 41

What is number needed to harm (NNH)?

How do you calculate NNH?

What does a NNH of 1 mean?

Answer 41

The number needed to harm (NNH) is the number of people who must be exposed to a treatment to cause harm to person who otherwise would not have been harmed.

To calculate NNH, the absolute risk increase (ARI) between the treatment and control groups must be known: NNH = 1 / ARI.

The lower the NNH, the more risk of harm; an NNH of 1 means that every patient treated is harmed.

Question 42

What bias are case-controls studies that rely on questionnaires or interviews to determine exposure status particulary susceptible to?

Answer 42

Case-control studies that rely on questionnaires or interviews to determine exposure status are particularly susceptible to misclassification bias in the form of recall bias. Subjects who have experienced an adverse event such as head and neck squamous cell carcinoma (HNSCC) are more likely to recall previous potential exposures (eg, pesticide exposure) than subjects who have not experienced an adverse event.

Question 43

What is the NNT?

How is it calculated?

Answer 43

The number needed to treat (NNT) is the number of patients who need to be treated with a specific treatment to avoid an additional negative event. NNT is the inverse of the absolute risk reduction.

Question 44

How does area under a curve (AUC) correlate to the accuracy of a test?

Answer 44

The accuracy of screening or diagnostic tests is quantified by the area under the ROC curve (AUC). The more accurate iS the test is (ie, higher sensitivity and specificity), the closer the AUC value is to 1.0. Tests with higher AUCs are more accurate than tests with lower AUCs.

Question 45

When does confounding occur?

What is used in case-control studies to control confounding?

Answer 45

Confounding occurs when the exposure-disease relationship is muddled by the effect 01 an extraneous factor associated with both exposure and disease. Confounding bias can result in the false association of an exposure with a disease.

Matching is used in case-control studies in order to control confounding. Matching variables should always be the potential confounders of the study (eg, age, race). Cases and controls are then selected based on the matching variables so that both groups have a similar distribution in accordance in with the variables.

Question 46

What is the purpose of randomization?

Answer 46

Randomization refers to the process of using random methods to assign subjects to experimental groups. Its purpose is to make experimental groups as similar as possible (except for the treatment assignment) to ensure that any difference observed between the groups is due exclusively to the treatment and not to other underlying factors.

Question 47

What creates the potential for attrition bias (selection bias) in prospective studies?

Answer 47

In prospective studies, disproportionate loss to follow-up between the exposed and unexposed groups creates the potential for attrition bias, a which is a form of selection bias. As a result, investigators generally try to achieve high patient follow-up rates in prospective studies.

Question 48

How does a sample size influence the power of a study?

Answer 48

A study’s power increases as its sample size increases. Therefore, the larger the sample, the greater the ability of a study to detect a difference when one truly exists.

Question 49

What are the main differences between the 2 types of study designs?

Answer 49

In observational designs (cross-sectional, case control, cohort), the researcher observes the effect of naturally occurring risk factors/exposures on outcomes of interest.

In experimental designs (randomized control trials, factorial, crossover), the researcher randomly assigns interventions to potential participants to assess the effect of the controlled interventions.

Question 50

How are scatter plots interpreted for correlation coefficient (r)?

Answer 50

Scatter plots are useful for crude data analysis. If a linear association is present between 2 variables, a correlation coefficient (r) mathematically describes how well a “line of best fit” (blue line in Figure) would correspond to the data points plotted. The value of r ranges from-1 to +1 and describes 2 important characteristics of an association: the strength and the polarity. The closer the r value is to its margins [-1, 1], the stronger the association.

Question 51

What are the two groups for case control studies?

Answer 51

Diseased cases vs Nondiseases cases and then compare risk factor frequency

Question 52

How can the attributable risk percent (ARP) in the exposed be derived?

Answer 52

Question 53

When are the results statistically significant based on a confidence interval (CI)?

Answer 53

CIs give a range of plausible values for an unknown parameter (eg, difference between 2 mean SBPs) based on results a from a sample. If the CI does not include the null value, then the result is statistically significant; if it crosses the null value, then the result is not statistically significant. All CIs have a null value, but the null value is not the same for all CIs. For an odds ratio or a relative risk (RR), the null value is 1 because these statistics are ratios (ie, RR = 1 represents no difference in risk between the groups). However, if the parameter of interest is a difference (eg, a difference in mean SBP between cocoa intake and control groups), then the null value is 0 because that represents no difference between the groups.

Question 54

What kind of study simultaneously measures exposures and outcomes?

Answer 54

A cross-sectional study (also known as a prevalence study) simultaneously measures exposures and outcomes. The cross-sectional study has a “snapshot” design that is frequently used in surveys, mostly because it is inexpensive and easy to perform. In this example, a snapshot was obtained of individuals randomly selected from the population; their blood samples were analyzed for the presence of the sodium channel protein mutation and the prevalence of hypertension was calculated. The subjects’ blood pressure was measured over 7 days to obtain an average measurement (likely to limit the results being impacted by white-coat hypertension and other transient causes of hypertension). The major limitation of a cross-sectional study design is that a temporal relationship between exposure and outcome is not always clear. However, in this case, demonstrating a temporal relationship was straightforward because the possession of a specific genotype clearly precedes hypertension.

Question 55

What is the main purpose of blinding in a research study?

Answer 55

Blinding technique is commonly used in clinical trials. The blinding can involve patients exclusively or both patients and physicians (double blinding). The main purpose of blinding is to prevent patient or researcher expectancy from interfering with the determination of an outcome. For example, a researcher’s belief in a positive outcome in treated patients can potentially result in observer bias.

Question 56

How is risk calculated?

Answer 56

Risk is the probability of developing a disease over a certain period of time. To calculate this probability, divide the number of affected subjects by the total number of subjects in the corresponding exposure group.

Question 57

When is regression analysis used?

Answer 57

Regression analysis is a statistica technique used to describe the elect th Or more independent variables (eg, exposures, risk factors), which may be quantitative or qualitative, can have on 1 quantitative dependent variable (ie, outcome).

Question 58

How do you interpret confidence interval of a population mean?

Answer 58