Ethics & Biostats

Question 1

Adverse events are estimated to occur in 1% of pediatric hospitalizations in the US annually, of which ___% could be avoided or prevented.

Answer 1

Adverse events are estimated to occur in 1% of pediatric hospitalizations in the US annually, of which 60% could be avoided or prevented.

Question 2

Langley Model for Improvement
The advantage of this framework is that is allows for multiple, rapid ____ cycles that can ultimately lead to quality improvement in a relatively brief time.

Answer 2

Langley Model for Improvement
The advantage of this framework is that is allows for multiple, rapid PDSA cycles that can ultimately lead to quality improvement in a relatively brief time.

Question 3

The Lean Methodology is a Quality Improvement model that focuses on_____

Lean makes use of the ____ (as opposed to consultants) to define the problems and find solutions.

Answer 3

The Lean Methodology is a Quality Improvement model that focuses on reducing waste in a process.

Lean makes use of the work team (as opposed to consultants) to define the problems and find solutions.

Question 4

Six Sigma (SS) focuses on reducing _____ in processes, and not necessarily waste. There are 2 types of variations in a process

When a process achieves six sigma it means there are only ____ defects (ie errors in health care) per million opportunities.

Answer 4

Six Sigma (SS) focuses on reducing variation in processes, and not necessarily waste. There are 2 types of variations in a process

When a process achieves six sigma it means there are only 3.4 defects (ie errors in health care) per million opportunities.

Question 5

Primary prevention

Answer 5

Primary prevention
Action taken before patient develops the disease and is targeted at preventing occurrence of the disease itself
Ex: lifestyle and habits

Question 6

Secondary prevention

Answer 6

Secondary prevention
Action that attempts to halt the progression of a disease at its initial stage before irreversible pathological changes take place, preventing complications
Ex: Statin

Question 7

Tertiary prevention

Answer 7

Tertiary prevention
Disease process has advanced beyond the early stages. Defined as taking all actions available to limit impairments and disabilities
Ex: cardiac rehabilitation and revascularization

Question 8

Quaternary prevention

Answer 8

Quaternary prevention
Set of health activities that might mitigate and/or limit the consequences of unnecessary or excessive intervention by the health system.
Ex: use EMR to limit unnecessary repeat cardiac cath procedures

Question 9

CASE-CONTROL study

Answer 9

CASE-CONTROL study
Selecting patients with a particular disease (cases) and patients without that disease (controls) and then retrospectively determining their exposure status

Question 10

Cohort studies

Answer 10

Cohort studies
Follow a population of patients with a disease over time. This approach is likely to yield important information about the natural hx of a disease.

Advantages of cohort studies:
1) Good for assessing the incidence of a condition
2) Predictor measurements are not affected by knowledge of the outcome since the outcome hasn’t occurred yet
3) Helpful for identifying possible predictor variables when few are known in a particular disease
Study design most likely to yield info about the prognosis of a condition

Disadvantages of cohort studies:

1) Difficult to infer causal relationships between predictor variables and the outcome (“correlation does not equal causation”)
2) They are less useful for rare outcomes.

Question 11

Ascertainment (sampling) bias

Answer 11

Ascertainment (sampling) bias: study population differs from target population due to nonrandom selection methods
Occurs when the results of a clinical study are distorted by knowledge of which intervention by the participants are assigned to.

Question 12

Publication bias

A ___ plot is helpful in assessing PUBLICATION BIAS.

Answer 12

Publication bias - trials with significant positive results are published but trials with negative/null results are not

A funnel plot is helpful in assessing PUBLICATION BIAS. Asymmetry in a funnel plot suggests publication bias.

Question 13

Effect modification (interaction bias)

Answer 13

Effect modification (interaction bias)
Results when an external variable (effect modifier) has a positive or negative impact on the observed effect of a risk factor (exposure) on disease status (outcome)

Extraneous variable changes the direction or strength of the effect that the independent variable (exposure or treatment) has on the dependent variable (outcome)

Stratification based on the modifier (eg analyzing the data by age) can help detect effect modification as it typically shows different effects in each stratum (eg children, adults). As a result, separate measures of outcome should be reported for each stratum

Stratified analysis can help distinguish between effect modification and confounding. With confounding, there is usually no significant difference seen with stratification.

Question 14

Length-time bias

Answer 14

Length-time bias occurs when the survival benefits of a screening test are overstated due to the detection of a disproportionate number of slowly progressive, benign cases.

Screening test detects less aggressive forms of a disease and therefore increases the apparent survival time

Results with subjects with a rapidly progressive form of disease are less likely to be detected by screening compared to those with slowly progressive diseases. Patients with slowly progressive disease tend to be asymptomatic for a longer period, increasing the likelihood that they will be diagnosed based on screening rather than clinical symptom.

Question 15

Lead time

Answer 15

Lead time bias occurs when a test diagnoses a disease earlier and as a result, the time from diagnosis until death appears prolonged even though there actually is no improvement in survival. This is different from length-time bias, which identifies more benign cases.

Question 16

In a normal distribution, ___% of all values are within 1 standard deviation from the mean; __% of all value are within 2 SDs from the mean, and ___% of all values are within 3 SDs from the mean.

Answer 16

In a normal distribution, 68% of all values are within 1 standard deviation from the mean; 95% of all value are within 2 SDs from the mean, and 99.7% of all values are within 3 SDs from the mean.

Question 17

For normally distributed continuous data, the ____ is the most valid choice for central location. For skewed distributions, the ____ is a more representative and valid measure of central location .

A negatively skewed distribution has >1 small-valued outliers and mean < median < mode

A positive skewed distribution has >1 large-valued outliers and mean > median > mode

Answer 17

For normally distributed continuous data, the mean is the most valid choice for central location. For skewed distributions, the MEDIAN is a more representative and valid measure of central location than the mean

A negatively skewed distribution has >1 small-valued outliers and mean < median < mode

A positive skewed distribution has >1 large-valued outliers and mean > median > mode

Question 18

The coefficient of determination = square the _____ (0.8 x 0.8). This expresses the percentage of the variability in the outcome factor that is explained by the predictor factor.

Answer 18

The coefficient of determination = square the correlation coefficient (0.8 x 0.8). This expresses the percentage of the variability in the outcome factor that is explained by the predictor factor.

Question 19

Correlation coefficient (r) values range from -1 to 1. The sign indicates a positive or negative direction of linear association between 2 variables; the null value is 0, which denotes no association. The closer the r value is to margins [-1, 1], the stronger the association. The correlation coefficient shows the strength and direction of linear association but does not imply causality.

Answer 19

Correlation coefficient (r) values range from -1 to 1. The sign indicates a positive or negative direction of linear association between 2 variables; the null value is 0, which denotes no association. The closer the r value is to margins [-1, 1], the stronger the association. The correlation coefficient shows the strength and direction of linear association but does not imply causality.

Question 20

Sensitivity

How often a ____ result correctly identifies those who ____ the disease.

A negative result on a highly sensitive diagnostic test helps to rule out a diagnosis (SnNout) bc the false-Negative rate is low. Highly sensitive tests are the most likely to detect a disease and the least likely to give false negative results. They are the most appropriate for screening

Sensitivity = ____ / _____

Does not change with prevalence of disease

Answer 20

Sensitivity

How often a positive result correctly identifies those who have the disease.

A negative result on a highly sensitive diagnostic test helps to rule out a diagnosis (SnNout) bc the false-Negative rate is low. Highly sensitive tests are the most likely to detect a disease and the least likely to give false negative results. They are the most appropriate for screening

Sensitivity = true positives / (true positives + false negatives) = TP / (# diseased)

Does not change with prevalence of disease

Question 21

Specificity

How often a _____ result correctly identifies those who do not have the disease. The higher the specificity, the lower the number of patients without disease that are testing falsely positive (the more certain you are that, given a negative result, the pt does not have the disease)

Highly specific tests are useful for confirmation. SPIN: help rule IN disease because the false-Positive rate is low

Specificity = ____ / ____

Does not change with prevalence of disease

Answer 21

Specificity

How often a negative result correctly identifies those who do not have the disease. The higher the specificity, the lower the number of patients without disease that are testing falsely positive (the more certain you are that, given a negative result, the pt does not have the disease)

Highly specific tests are useful for confirmation. SPIN: help rule IN disease because the false-Positive rate is low

Specificity = true negatives / (true negatives + false positives) = TN / (# Not diseased)

Does not change with prevalence of disease

Question 22

Positive predictive value

The probability that disease is present given a positive result.

PPV =____ / ____

As _____ increases, the positive predictive value also increases and the number of false positives decreases. The test with the HIGH _____ is the best choice as it will most likely also have the highest PPV. Higher specificity is associated with fewer FPs and increase tthe PPV (high sensitivity means fewer FNs)

Dependent on ____ of the disease in the tested population

Answer 22

Positive predictive value

The probability that disease is present given a positive result.

PPV = true positives / (true positives + false positives)

As specificity increases, the positive predictive value also increases and the number of false positives decreases. The test with the HIGH SPECIFICITY is the best choice as it will most likely also have the highest PPV. Higher specificity is associated with fewer FPs and increase tthe PPV (high sensitivity means fewer FNs)

Dependent on prevalence of the disease in the tested population

Question 23

Negative predictive value
The probability that disease is absent given a negative result.

NPV = ____ / ___

Dependent on ____ of the disease in the tested population

Answer 23

Negative predictive value
The probability that disease is absent given a negative result.

NPV = true negative / (true negative + false negatives)

Dependent on prevalence of the disease in the tested population

Question 24

Positive likelihood ratio
A ratio representing the likelihood of ____

Probability of pt w disease testing positive divided by probable of a pt without the disease testing positive
= true positive rate / false positive rate

LR+ = ____ / ____

LRs do not change with prevalence of the disease

Ratio > 10 = strong evidence to rule in the disease

LR >1 suggests disease presence; the higher the LR, the more likely the disease presence. LR <1 argues against the disease; the lower the LR, the less likely the disease presence

0.5-2 = no evidence to rule in or rule out the disease

<0.1 = strong evidence to rule out the diseasE

Answer 24

Positive likelihood ratio
A ratio representing the likelihood of having the disease given a positive result.

Probability of pt w disease testing positive divided by probable of a pt without the disease testing positive
= true positive rate / false positive rate

LR+ = sensitivity / (1 - specificity)

LRs do not change with prevalence of the disease

Ratio > 10 = strong evidence to rule in the disease

LR >1 suggests disease presence; the higher the LR, the more likely the disease presence. LR <1 argues against the disease; the lower the LR, the less likely the disease presence

0.5-2 = no evidence to rule in or rule out the disease

<0.1 = strong evidence to rule out the diseasE

Question 25

Negative likelihood ratio
A ratio representing the likelihood of ___.
Probability of a pt with the disease testing negative divided by the probability of a pt without the disease testing negative.
= false negative rate / true negative rate

LR - = ____ / ____

Does not change with prevalence of disease

Answer 25

Negative likelihood ratio
A ratio representing the likelihood of having the disease given a negative result.
Probability of a pt with the disease testing negative divided by the probability of a pt without the disease testing negative.
= false negative rate / true negative rate

LR - = (1 - sensitivity) / specificity

Does not change with prevalence of disease

Question 26

____ normogram can be used to determine how a test with a known ____ ratio can predict the ____ probability if the pretest probability is known. This can be done by drawing a straight line starting from the pretest probability through the likelihood ratio and recording the result of the post-test probability. The pretest probability can be estimated based on disease prevalence.

Answer 26

Fagan normogram can be used to determine how a test with a known likelihood ratio can predict the post-test probability if the pretest probability is known. This can be done by drawing a straight line starting from the pretest probability through the likelihood ratio and recording the result of the post-test probability. The pretest probability can be estimated based on disease prevalence.

Question 27

A screening test must have a high _______. This high sensitivity helps to ‘RULE OUT’ the disease, giving as few false-negative results as possible (SNOut). Furthermore, the high sensitivity increases the negative predictive value (NPV) of the test:

Answer 27

A screening test must have a high sensitivity. This high sensitivity helps to ‘RULE OUT’ the disease, giving as few false-negative results as possible (SNOut). Furthermore, the high sensitivity increases the negative predictive value (NPV) of the test:

Question 28

Crude mortality rate = number of deaths divided by total population size

Cause-specific mortality rate = # deaths from particular disease divided by total population size

CASE FATALITY RATE refers to the proportion of people with a particular condition who end up dying from the condition. It is different from mortality rate, which describes the general population’s likelihood of dying from the disease.
Case fatality rate = # deaths from a specific disease divided by the number of people affected by the disease

Standard mortality rate (SMR) = observed number of deaths in population of interest/expected number of deaths from reference population (standard)

Standard mortality ratio quantifies mortality in a specific group as compared to the general population. It is the ratio of observed to expected number of deaths in a specific group of the general population under the assumption that mortality rates for the group are the same as those for the general population. The expected number of deaths is calculated based on age-specific mortality rates in a standard reference population. This value is then contrasted to the number of deaths in the specific group: SMR = observed number of deaths / expected number of deaths.
SMR <1 indicates that the number of observed deaths in the study group is lower than what is expected
SMR = 1 indicates that the number of observed deaths in the study group is equal to what is expected
SMR >1 indicates that the number of observed deaths in the study group is greater than what is expected
A confidence interval CI that does not include the null value (1.0 for SMRs) indicates a statistically significant difference between the observed and expected number of deaths.

Answer 28

Crude mortality rate = number of deaths divided by total population size

Cause-specific mortality rate = # deaths from particular disease divided by total population size

CASE FATALITY RATE refers to the proportion of people with a particular condition who end up dying from the condition. It is different from mortality rate, which describes the general population’s likelihood of dying from the disease.
Case fatality rate = # deaths from a specific disease divided by the number of people affected by the disease

Standard mortality rate (SMR) = observed number of deaths in population of interest/expected number of deaths from reference population (standard)

Standard mortality ratio quantifies mortality in a specific group as compared to the general population. It is the ratio of observed to expected number of deaths in a specific group of the general population under the assumption that mortality rates for the group are the same as those for the general population. The expected number of deaths is calculated based on age-specific mortality rates in a standard reference population. This value is then contrasted to the number of deaths in the specific group: SMR = observed number of deaths / expected number of deaths.
SMR <1 indicates that the number of observed deaths in the study group is lower than what is expected
SMR = 1 indicates that the number of observed deaths in the study group is equal to what is expected
SMR >1 indicates that the number of observed deaths in the study group is greater than what is expected
A confidence interval CI that does not include the null value (1.0 for SMRs) indicates a statistically significant difference between the observed and expected number of deaths.

Question 29

Odds ratio

Measure of association between an exposure and an outcome.

The odds that an outcome/event will occur in the presence of a particular exposure (eg therapy) compared to the odds of that outcome in a control group

Odds ratio is a measure of association commonly used in ______ studies, where it represents the odds that a case was exposed divided by the odds that a control was exposed

OR = ___/___

OR > 1 means that the exposure is associated with higher odds of the outcome
OR <1 means that the exposure is associated with lower odds of the outcome (protective)
OR =1 means that the exposure is not associated with the outcome/disease

Answer 29

Odds ratio

Measure of association between an exposure and an outcome.

The odds that an outcome/event will occur in the presence of a particular exposure (eg therapy) compared to the odds of that outcome in a control group

Odds ratio is a measure of association commonly used in case-control studies, where it represents the odds that a case was exposed divided by the odds that a control was exposed

OR = ad/bc

OR > 1 means that the exposure is associated with higher odds of the outcome
OR <1 means that the exposure is associated with lower odds of the outcome (protective)
OR =1 means that the exposure is not associated with the outcome/disease

Question 30

Relative risk

Represents a measure of outcome in follow-up studies. It is the risk of the outcome in the exposed group divided by the risk of the outcome in the unexposed group.

RR =____ / ____

RR = treatment rate / control rate

RR <1.0 indicates decreased risk in the group in the numerator
RR = 1.0 indicates no difference in risk between the groups (confidence interval must exclude the null value if it is statistically significant)
RR >1.0 indicates increased risk in the group in the numerator

If confidence of an estimate reflects lack of statistical significant, then the corresponding p-value should also reflect the same conclusion. Therefore, if the 95% CI includes the null value, the p-value should be >0.05 (not statistically significant). Conversely, if the 95% CI does not include the null value, the p-value should be <0.05 (statistically significant).

Answer 30

Relative risk

Represents a measure of outcome in follow-up studies. It is the risk of the outcome in the exposed group divided by the risk of the outcome in the unexposed group.

RR = risk of disease in exposed group / risk of disease in unexposed group

RR = treatment rate / control rate

RR <1.0 indicates decreased risk in the group in the numerator
RR = 1.0 indicates no difference in risk between the groups (confidence interval must exclude the null value if it is statistically significant)
RR >1.0 indicates increased risk in the group in the numerator

If confidence of an estimate reflects lack of statistical significant, then the corresponding p-value should also reflect the same conclusion. Therefore, if the 95% CI includes the null value, the p-value should be >0.05 (not statistically significant). Conversely, if the 95% CI does not include the null value, the p-value should be <0.05 (statistically significant).

Question 31

Hazard ratio

Likelihood of an event occurring in a ___ group relative to the ___ group

The null value for HR is 1.0

HR<1.0 indicates an event is less likely to occur in a treatment group than the control group, indicates a protective effect. HR >1.00 indicates a detrimental effect. HR=1.00 means there is no difference in risk between the 2 groups.

HR is similar to relative risk RR, except that RR is usually calculated at the end of a study (or other defined endpoint) to convey the risk of an event occurring within that time frame. In contrast, HRs are a measure of the instantaneous risk of an event occurring, usually during a subset of the total study period. HRs differ from RR in that they can be calculated at multiple time intervals throughout a study period.

Answer 31

Hazard ratio

Likelihood of an event occurring in a treatment group relative to the control group

The null value for HR is 1.0

HR<1.0 indicates an event is less likely to occur in a treatment group than the control group, indicates a protective effect. HR >1.00 indicates a detrimental effect. HR=1.00 means there is no difference in risk between the 2 groups.

HR is similar to relative risk RR, except that RR is usually calculated at the end of a study (or other defined endpoint) to convey the risk of an event occurring within that time frame. In contrast, HRs are a measure of the instantaneous risk of an event occurring, usually during a subset of the total study period. HRs differ from RR in that they can be calculated at multiple time intervals throughout a study period.

Question 32

Relative risk reduction

Quantifies the proportion of risk reduction attributable to a specific intervention or exposure as compared to a control. RRR considers the risk for disease in the exposed/intervention group and the unexposed/control group as follows:

RRR = (risk in unexposed - risk in exposed) / (risk in unexposed)

RRR = ____

Answer 32

Relative risk reduction

Quantifies the proportion of risk reduction attributable to a specific intervention or exposure as compared to a control. RRR considers the risk for disease in the exposed/intervention group and the unexposed/control group as follows:

RRR = (risk in unexposed - risk in exposed) / (risk in unexposed)

RRR = 1 - RR

Question 33

Attributable risk percent

Measure of excess risk and estimates the proportion of disease among exposed subjects that is attributed to exposure status

2 approaches to calculate ARP

• ARP = (risk in exposed - risk in unexposed) / risk in exposed
• ARP = ______

Answer 33

Attributable risk percent

Measure of excess risk and estimates the proportion of disease among exposed subjects that is attributed to exposure status

2 approaches to calculate ARP

• ARP = (risk in exposed - risk in unexposed) / risk in exposed
• ARP = (RR - 1) / RR

Question 34

Intention to treat analysis

Every subject is analyzed according to his or her randomized group assignment, regardless of changes that may occur after randomization, which maintains similarity in treatment groups. “Once randomized, always randomized.” Intention to treat analysis preserves sample size and statistical power is therefore maintained. ITT minimizes _____, or the incorrect rejection of a true null hypothesis.

Intention to treat includes data from all patients who were randomly assigned to a group even if they did not complete the study or switched groups.

PRESERVES THE BENEFITS OF ____ AND PREVENT BIAS DUE TO SELECTIVE NON-COMPLIANCE

Answer 34

Intention to treat analysis

Every subject is analyzed according to his or her randomized group assignment, regardless of changes that may occur after randomization, which maintains similarity in treatment groups. “Once randomized, always randomized.” Intention to treat analysis preserves sample size and statistical power is therefore maintained. ITT minimizes type I error, or the incorrect rejection of a true null hypothesis.

Intention to treat includes data from all patients who were randomly assigned to a group even if they did not complete the study or switched groups.

PRESERVES THE BENEFITS OF RANDOMIZATION AND PREVENT BIAS DUE TO SELECTIVE NON-COMPLIANCE

Question 35

Per protocol analysis

Only those who complete the study are included in the final analysis
It is more likely than intention to treat analysis to be affected by bias, since the act of dropping out or switching groups is usually non-random.

Answer 35

Per protocol analysis

Only those who complete the study are included in the final analysis
It is more likely than intention to treat analysis to be affected by bias, since the act of dropping out or switching groups is usually non-random.

Question 36

When confidence intervals of means or proportions for >2 groups do not overlap, it is always correct to establish a statistically significant difference between the groups. However, when confidence intervals overlap, there may or may not be a statistically significant difference between the groups.

A confidence interval that crosses the null value or a p-value > alpha error cutoff (typically 0.05) denotes a result that is not statistically significant.

The width of the CI is inversely related to_____: increasing the sample size decreases the CI, indicating higher precision of the dataset.

Answer 36

When confidence intervals of means or proportions for >2 groups do not overlap, it is always correct to establish a statistically significant difference between the groups. However, when confidence intervals overlap, there may or may not be a statistically significant difference between the groups.

A confidence interval that crosses the null value or a p-value > alpha error cutoff (typically 0.05) denotes a result that is not statistically significant.

The width of the CI is inversely related to sample size: increasing the sample size decreases the CI, indicating higher precision of the dataset.

Question 37

Number needed to harm

Measure that indicates how many patients need to be exposed to a particular risk factor over a specific period before a harmful event occurs in 1 patient

NNH = __ / ___

Answer 37

Number needed to harm

Measure that indicates how many patients need to be exposed to a particular risk factor over a specific period before a harmful event occurs in 1 patient

NNH = 1 / absolute risk increase (or attributable risk)

Question 38

Number needed to treat

Number of patients who need to be treated to prevent 1 bad outcome

NNR = ___/ ____

Helps providers determine the risk-benefit ratio for an individual patient for a specific therapy.

Answer 38

Number needed to treat

Number of patients who need to be treated to prevent 1 bad outcome

NNR = 1/ absolute risk reduction (ARR) = 1 / (control event rate - experimental event rate)

Helps providers determine the risk-benefit ratio for an individual patient for a specific therapy.

Question 39

Absolute risk reduction

The difference in the risk of adverse outcomes between the study population and a control group

Percentage indicating the actual difference in event rate between control and treatment groups

ARR = ____ - ____

Answer 39

Absolute risk reduction

The difference in the risk of adverse outcomes between the study population and a control group

Percentage indicating the actual difference in event rate between control and treatment groups

ARR = control group event rate CER - experimental group event rate EER

Question 40

Absolute risk (AR) = (patients with event in group)/(total pts in group)

Answer 40

Absolute risk (AR) 
Absolute risk (AR) = (patients with event in group)/(total pts in group)

Question 41

The McNemar test compares the difference between 2 paired ____; patients serve as their own control (before and after treatment in the same subjects).

Answer 41

The McNemar test compares the difference between 2 paired proportions; patients serve as their own control (before and after treatment in the same subjects).

Question 42

Kappa statistic is a quantitative measure of inter-rater _____ (inter-rater concordance).

Reflects the extent to which inter-rater agreement represents an improvement on chance agreement alone.

Kappa values range from -1 (perfect disagreement) to +1 (perfect agreement), with kappa = 0 suggesting agreement due to chance, kappa <0 suggesting less than chance agreement (possibly intentional disagreement), and kappa >1 suggesting greater than chance agreement

Following arbitrary cutoffs are often used for kappa: 0-0.2 represents negligible improvement over chance agreement, 0.21-0.4 is minimal, 0.41-0.6 is fair, 0.61-0.8 is good, and >0.8 is excellent

Answer 42

Kappa statistic is a quantitative measure of inter-rater RELIABILITY (inter-rater concordance).

Reflects the extent to which inter-rater agreement represents an improvement on chance agreement alone.

Kappa values range from -1 (perfect disagreement) to +1 (perfect agreement), with kappa = 0 suggesting agreement due to chance, kappa <0 suggesting less than chance agreement (possibly intentional disagreement), and kappa >1 suggesting greater than chance agreement

Following arbitrary cutoffs are often used for kappa: 0-0.2 represents negligible improvement over chance agreement, 0.21-0.4 is minimal, 0.41-0.6 is fair, 0.61-0.8 is good, and >0.8 is excellent

Question 43

Power of a study

Def: the probability to detect the difference in the outcome of interest between 2 groups, if such a difference exists

Power is the likelihood of not committing a ____ (making a false negative decision). By increasing the power, the chance of making a Type 2 error decreases.
If the power is 80%, the chance of a type 2 error is 20%. The only way to increase the power is to use more subjects, to continue for a longer duration, or both

Power = \_\_ - \_\_
B = probability of making a type 2 error (failure to reject null hypothesis when it is false)

Increasing the ____ increases the power of a study and makes the confidence interval narrower

Increasing the power of the study will allow p value to reach statistical significance

Power depends on sample size (larger size increases power), effect size (and standard deviation), and alpha and beta error levels.

Answer 43

Power of a study

Def: the probability to detect the difference in the outcome of interest between 2 groups, if such a difference exists

Power is the likelihood of not committing a Type 2 error (making a false negative decision). By increasing the power, the chance of making a Type 2 error decreases.
If the power is 80%, the chance of a type 2 error is 20%. The only way to increase the power is to use more subjects, to continue for a longer duration, or both

Power = 1 - B
B = probability of making a type 2 error (failure to reject null hypothesis when it is false)

Increasing the sample size increases the power of a study and makes the confidence interval narrower

Increasing the power of the study will allow p value to reach statistical significance

Power depends on sample size (larger size increases power), effect size (and standard deviation), and alpha and beta error levels.

Question 44

Meta-analysis

Def: Pooling the data from several studies to perform an analysis is called meta-analysis.

Is a valuable epidemiologic tool used to increase the ____ of a study.

It is the responsibility of the author of the meta-analysis to ensure that the individual studies are as similar as possible, that those studies are of high quality, and that there is minimal publication bias. The quality of the meta-analysis is no better than the quality of the individual studies “garbage in, garbage out.”

Answer 44

Meta-analysis

Def: Pooling the data from several studies to perform an analysis is called meta-analysis.

Is a valuable epidemiologic tool used to increase the power of a study.

It is the responsibility of the author of the meta-analysis to ensure that the individual studies are as similar as possible, that those studies are of high quality, and that there is minimal publication bias. The quality of the meta-analysis is no better than the quality of the individual studies “garbage in, garbage out.”

Question 45

Type I error (alpha)

Concluding there is a ____ when ____(ie rejecting the null hypothesis incorrectly). Type 1 errors are exposed by investigating the ____ (statistical significance) of the results

Conducting multiple independent hypothesis tests without proper adjustment to the alpha level increases the rate of type I error. This means that, ____ INCREASES THE LIKELIHOOD OF A TYPE I ERROR. When evaluating multiple secondary endpoints, there is a higher probability of erroneously finding a statistically significant result w one of these endpoints due to chance alone. This phenomenon is known as the multiplicity, or multiple testing, problem.

Answer 45

Type I error (alpha)

Concluding there is a difference in outcomes when there is not (ie rejecting the null hypothesis incorrectly). Type 1 errors are exposed by investigating the p value (statistical significance) of the results

Conducting multiple independent hypothesis tests without proper adjustment to the alpha level increases the rate of type I error. This means that, TESTING FOR MULTIPLE SECONDARY ENDPOINTS INCREASES THE LIKELIHOOD OF A TYPE I ERROR. When evaluating multiple secondary endpoints, there is a higher probability of erroneously finding a statistically significant result w one of these endpoints due to chance alone. This phenomenon is known as the multiplicity, or multiple testing, problem.

Question 46

TYPE II ERROR (beta)

Concluding _____ when _____

The probability is inversely related to the ____ of a study (1-B). Studies with a larger sample size have greater power to detect differences, and are less likely for type 1 or type 2 errors to occur. Small studies have less power, increasing the chance of a type 2 error

Answer 46

TYPE II ERROR (beta)

Concluding there is no difference when there actually is (failing to reject the null hypothesis).

The probability is inversely related to the power of a study (1-B). Studies with a larger sample size have greater power to detect differences, and are less likely for type 1 or type 2 errors to occur. Small studies have less power, increasing the chance of a type 2 error

Question 47

Kaplan-Meier survival curve depicts the probability of ____ at various time points during the study. This probability is calculated based on the proportion of subjects who are alive at a given time. Depending on the number of study groups, >2 curves can be displayed simultaneously and compared. Survival of different groups can be compared by the log-rank test, which reports a p-value. In most studies, a p-value <0.05 is considered significant.

Answer 47

Kaplan-Meier survival curve depicts the probability of survival at various time points during the study. This probability is calculated based on the proportion of subjects who are alive at a given time. Depending on the number of study groups, >2 curves can be displayed simultaneously and compared. Survival of different groups can be compared by the log-rank test, which reports a p-value. In most studies, a p-value <0.05 is considered significant.

Question 48

A receiver-operating characteristic (ROC) curve for a given diagnostic test plots sensitivity on the y-axis and (1-specificity) on the x-axis. It can help determine the best cutoff point to use depending on the optimal desired parameters for sensitivity and specificity. The area under the curve of a ROC curve is a reflection of _____. A larger AUC means better discrimination and higher diagnostic accuracy. An ideal test would have an AUC of 1.0 (100% sensitivity and specificity) and a nondiscriminating test would have an AUC of 0.5. A test with no predictive value would be represented by a straight line.

Answer 48

A receiver-operating characteristic (ROC) curve for a given diagnostic test plots sensitivity on the y-axis and (1-specificity) on the x-axis. It can help determine the best cutoff point to use depending on the optimal desired parameters for sensitivity and specificity. The area under the curve of a ROC curve is a reflection of diagnostic accuracy. A larger AUC means better discrimination and higher diagnostic accuracy. An ideal test would have an AUC of 1.0 (100% sensitivity and specificity) and a nondiscriminating test would have an AUC of 0.5. A test with no predictive value would be represented by a straight line.