NAPLEX Questions: Biostatistics

The distortion of the central tendency caused by outliers is reduced by ___

Collecting more values

When the number of values is small, an outlier has a __ impact on the mean. The ___ should be used as the best indicator of a central tendency.

Large; median

__% of the values in a normal distribution fall within one SD of the mean, and __% of the values fall within two SDs of the mean.

68%; 95%

When data in a normal curve has more high values, the hump moves to the right, and there are more outliers to the left. The data is skewed to the ___.

Left

When data in a normal curve has more low values, the hump moves to the left, and there are more outliers to the right. The data is skewed to the __.

Right

Define the null hypothesis

Null means no! A null hypothesis states that there is no stastically significant difference between two groups. What the researcher does not want to be true.

Define the alternative hypothesis

The alternative hypothesis states that there is a statistically significant difference between two groups; what the researcher tries to prove.

Define the alpha level

Maximum permissible error margin. Commonly 5%.

If the p-value is less than or equal to alpha, the null hypothesis is ___

Rejected; result is statistically significant.

If the p-value is greater than alpha, the null hypothesis is ___

Accepted aka failed to be rejected; result is not statistically significant

How do you explain that the p-value is <0.05 in terms of probability/confidence interval?

There's a 95% probability the conclusion is correct and 5% chance that it's not.

When difference (means) data is compared, the data is significant when the CI range does not include ___.

Zero

Define a type 1 error

Rejecting the null hypothesis in error. Saying there is a difference when there isn't one. False positive.

Define a type 2 error

Accepting the null when it should have been rejected. Saying there was no difference when there actually was. Manufacturer's worst nightmare.

Define power

The probability that a test will reject the null hypothesis appropriately. No mistakes.

Power is affected by three things

1. Number of outcome values collected (more subjects)

2. The difference in outcome rates between the groups

3. The significance level

Define relative risk (risk ratio)

Ratio of risk in the exposed group divided by risk in the control group

Risk ratio = 1 implies __

No difference in risk of the outcome between the groups

Risk ratio >1 implies __

Greater risk of the outcome

Risk ratio <1 implies __

Lower risk (reduced risk) of the outcome

Relative risk reduction (RRR) definition

How much less likely the event risk is in the treatment group relative to the risk in the control group

RR and RRR are calculated as percentage or decimals?

Decimals unless they ask for percentages!

Absolute risk reduction definition

Absolute risk reduction includes the reduction in risk AND the incidence rate of the outcome. For example, it's great to have a 50% reduction in nausea risk, but what if the risk was only 0.1% in the first place?

Absolute risk reduction definition in terms of calculation

Treatment effect minus the placebo effect is the absolute risk reduction - the true incidence rate of the outcome

Number needed to treat definition; *round up or round down?*

Number of patients who need to be treated for a period of time in order for one patient to benefit. Round up to be more conservative.

Number needed to harm definition; *round up or round down?*

Number of patients who need to be treated for a certain period of time in order for one patient to be harmed. Round down to be more conservative.

Define odds ratio

OR is used to calculate the odds that an outcome will occur with an exposure compared to the odds that it will occur without the exposure. Done instead of RR and is better suited for a treatment in a case-control study. RR compares treatment vs. placebo. OR calculates exposure vs. no exposure.

Define hazard ratio

Calculated when analyzing death or disease progression. Rate at which unfavorable event occurs over a short period of time.

OR or HR = 1 means...

The event rate is the same in the treatment and control arms. No advantage to treatment.

OR or HR > 1 means...

The event rate in the treatment group is higher than the event rate in the control group. Ex: HR of 2 means that there were twice as many deaths in the treatment group

OR or HR <1 means...

The event rate in the treatment group is lower than the event rate in the control group. Ex: HR of 0.5 would indicate half as many deaths in the treatment group

__ and __ are two types of discrete data

Nominal and ordinal

Define nominal data and give an example

Data that has categories or names. Female or male.

Define ordinal data and give an example.

Data that has a logical order but is NOT equivalent between the numbers, such as a pain scale.

__ and __ are the two types of continuous data. They both have an order with steps that continuously increase by the same amount.

Interval and ratio

What is the difference between interval and ratio data?

Interval data has no meaningful zero (zero does NOT equal none). Example is temperature. Ratio data does have a meaningful zero. Example is heart rate. HR can equal zero (dead).

A t-test is used to assess significance in studies with __ data values

Continuous

If studies have both a treatment and a control group, you would use an independent or unpaired __ to evaluate.

T-test

If the study has three or more groups, you would use __ to evaluate.

ANOVA

A chi-square test is used to assess significance in studies with __ data values

Discrete

Define an independent variable

The variable that is manipulated by the researcher and has an effect on the dependent variable. Example is picking certain patients to be in treatment and control groups. Inclusion criteria (can anyone have HTN?)

Define a composite endpoint

Combines multiple, individual endpoints into one measurement. Must use the composite endpoint value when assessing the composite endpoint because just adding the individual endpoints will NOT equal the same thing

Correlation definition

Statistical technique that is used to determine if one variable is related to (or changes) another variable

___ is used to test correlation in ordinal, ranked data

Spearman's Rank Order Correlation, aka Rho

__ is used to test correlation in continuous data. It indicates the strength and direction of the relationship between two variables.

Pearson's correlation coefficient, aka r

Define regression

Regression is used to describe the relationship between a dependent variable and one or more independent variables, or how much the value of the dependent variable changes when the independent variable changes.

Name the three types of regression

1. Linear - continuous data

2. Logistic - categorical data

3. Cox regression - categorical data in a survival analysis

Intention-to-treat analysis definition

Analysis includes data for all patients originally allocated to each treatment group, even if the patient did not complete the trial according to the study protocol

Per protocol analysis definition

Conducted for the subset of the trial population who actually completed the study. Does not include drop outs.

Define the sensitivity of a lab test

How effectively a test identifies a specific condition. A test with 100% sensitivity will be positive in all patients with that condition. Can you tell me who does have the condition? Be sensitive when someone is HIV positive.

Define the specificity of a lab test

How effectively a test identifies patients without the conditions. A test with 100% specificity will be negative in all patients with the condition. Can you tell me who doesn't have the condition? Be specific when saying someone is HIV negative.

Confidence interval interpretation: comparing difference data. A result is statistically significant if the CI does not cross __

zero

Confidence interval interpretation: comparing ratio data (RR, OR, or HR). A result is statistically significant if the CI does not cross

one

Define a case-control study

Retrospectively compares patients with a disease (cases) to those without the disease (controls) to see if a relationship exists between the disease and various risk factors.

Define a cohort study

Prospectively or retrospectively (less common) compares outcomes of a group of patients exposed and not exposed to a treatment to see if they develop the outcome. Done when an intervention would be unethical.

Define a cross-sectional survey

Estimates the relationship between variables and outcomes (prevalence) at one particular time (cross-section) in a defined population. Can identify possible hypotheses.

Define case report and case series

Describe an adverse reaction or a unique condition that appears in a single patient (case report) or a few patients (case series).

Define a randomized controlled trial.

Compares an experimental treatment to a control (placebo or existing treatment) to determine which is better. Subjects with the desired characteristics are carefully selected to exclude patients with characteristics that may also influence outcome and then are randomized and sometimes blinded.

Define a parallel study

Subjects are randomized to the treatment or control arms for the entire study

Define a crossover study

Patients are randomized to one of two sequential treatments (A, then B, or B, then A)

Define a factorial study

Randomizes to more than the usual two groups to test a number of experimental conditions

Define a meta-analysis

Combines results from multiple studies in order to develop a conclusion that has greater statistical power than is possible from the individual smaller studies.

Define a systematic review article

Summary of the clinical literature that focuses on a specific topic or question. Begins with a question followed by a literature search and summary of information found.