NAPLEX Biostatistics Review

Flashcards covering biostatistics concepts from the NAPLEX Webinar, including bias types, data classifications, trial populations, and statistical test selection.

Performance Bias

A type of bias that occurs when study groups are not treated equally by investigators or when patients alter their behavior, often minimized by blinding.

Selection Bias

Bias that occurs prior to trial implementation during the baseline patient enrollment process.

Attrition Bias

Bias related to unequal dropout rates between the different groups in a study.

Publication Bias

Bias occurring after trial completion where only positive results are published.

Nominal Data

Categorical data with no inherent rank or magnitude, such as mortality (all-cause mortality) or marital status.

Ordinal Data

Categorical data with a ranked order, but where the clinical difference between ranks is not mathematically equal, such as NYHA Functional Class (I-IV).

Ratio Data

Continuous numerical data with an absolute zero, such as heart rate in beats per minute.

Type 1 Error

A false positive result, represented by the Greek letter $\alpha$.

Type 2 Error

A false negative result, represented by the Greek letter $\beta$.

Power

The ability of a study to detect a difference if one exists, which is influenced by sample size.

Positive (Right) Skew

A distribution where high-value outliers pull the tail of the data to the right; in this case, the median is typically the most accurate measure of central tendency.

Odds Ratio ($OR$)

An association measure comparing the odds of an outcome with exposure versus without, typically used in Case-Control studies.

Hazard Ratio ($HR$)

The risk of an event occurring at any given time, commonly used in survival analysis.

Analysis of Variance ($ANOVA$)

A statistical test used to compare continuous, parametric data across three or more independent groups.

Kruskal-Wallis Test

A statistical test for three or more independent groups using ordinal (non-parametric) data.

Chi-Square Test

A statistical test used specifically for nominal data.

Student's T-Test

A statistical test used for continuous data when comparing exactly two independent groups.

Intention-to-Treat ($ITT$) Analysis

An analysis population that includes all randomized patients, including dropouts, providing a conservative, real-world estimate of clinical utility.

Per-Protocol ($PP$) Analysis

An analysis population that only includes patients who completed the study and adhered strictly to the protocol.

Forest Plot Diamond

Represents the confidence interval of the pooled result in a meta-analysis; if it does not cross the line of no effect, the result is statistically significant.

Noninferiority Margin ($\Delta$)

A clinical threshold (Delta) used to prove a drug is no worse than the standard; the entire 95% Confidence Interval must sit to the right (above) of this margin to prove noninferiority.

Meta-Analyses

Considered the strongest hierarchy of evidence, consisting of systematic reviews of existing studies.

Case-Control Study

A retrospective study that compares people who have a disease to people who do not have the disease.

Cohort Study

A study that follows a group to compare those with a specific exposure to those without that exposure.

Number Needed to Treat ($NNT$)

The number of patients who need to be treated to prevent one additional event, calculated as $\frac{1}{ARR}$, always rounded up to the nearest whole number.

Absolute Risk Reduction ($ARR$)

The absolute difference in the risk of an event between the placebo group and the treatment group.