How is Relative Risk (RR) calculated from a Cohort study?

RR = Risk in exposed / Risk in unexposed = [a / (a + b)] / [c / (c + d)]. If RR = 1, risk is identical; RR > 1 indicates increased risk; RR < 1 indicates protective association.

How is Absolute Risk Reduction (ARR) and Number Needed to Treat (NNT) calculated?

ARR = Event rate in control group (CER) - Event rate in treatment group (EER). NNT = 1 / ARR (always rounded UP to the nearest whole integer).

How is the Odds Ratio (OR) calculated from a 2x2 table in a Case-Control study?

OR = (a * d) / (b * c), where 'a' = exposed cases, 'b' = exposed controls, 'c' = unexposed cases, and 'd' = unexposed controls. Approximates Relative Risk when disease prevalence is low ('rare disease assumption').

Can an adult Jehovah's Witness patient with full decision-making capacity refuse life-saving blood transfusion?

Yes. Autonomy empowers a competent adult patient to refuse any medical treatment, even if that refusal results in inevitable death. The physician must verify capacity and explore acceptable alternatives (cell saver, erythropoietin), but cannot override their refusal.

Under what circumstances may a physician breach patient confidentiality without patient consent?

1. Reportable infectious diseases (e.g., tuberculosis, syphilis, measles); 2. Suspected child abuse or elder abuse (mandatory reporting); 3. Clear, imminent, serious threat of self-harm or violence against an identifiable third party ('Tarasoff duty to warn'); 4. Impaired drivers unfit to operate vehicles.

How are Sensitivity and Specificity defined and calculated from a 2x2 contingency table?

Sensitivity = TP / (TP + FN); the proportion of true disease cases correctly detected by the test. Specificity = TN / (TN + FP); the proportion of non-diseased individuals correctly identified as negative.

How does Decision-Making Capacity differ from Competence?

Capacity is a clinical determination made by a healthcare provider for a specific medical decision at a specific point in time. Competence is a legal determination made by a judge in a court of law.

What is Confounding and how can it be controlled during study design and statistical analysis?

A third variable associated with both exposure and outcome, but not on the causal pathway. Controlled in design by Randomization, Matching, and Restriction; controlled in analysis by Stratification and Multivariable Regression.

How does changing disease prevalence affect Sensitivity, Specificity, Positive Predictive Value (PPV), and Negative Predictive Value (NPV)?

Sensitivity and Specificity are intrinsic test properties and do NOT change with prevalence. Higher prevalence increases PPV and decreases NPV; lower prevalence decreases PPV and increases NPV.

What is Length-Time Bias in cancer screening programs?

Screening tests disproportionately detect slowly progressing, indolent cases with long asymptomatic phases, while missing rapidly progressive aggressive cancers that present symptomatically between screening intervals.

What is Recall Bias and in which study design is it most commonly encountered?

A form of measurement/information bias where cases recall past exposures differently (often more thoroughly or falsely) than healthy controls; classically plagues retrospective Case-Control studies.

How is Number Needed to Harm (NNH) calculated?

NNH = 1 / Attributable Risk (where Attributable Risk = Adverse event rate in exposed - Adverse event rate in unexposed). Always rounded DOWN to the nearest integer for conservative harm estimation.

Under what conditions can a minor receive medical treatment without parental knowledge or consent ('emancipated minor' vs. specific exceptions)?

Emancipation (married, self-supporting/military, or court decree) grants full consent rights. Regardless of emancipation, most jurisdictions allow confidential minor consent for: 1. Sexually transmitted infection treatment; 2. Contraception/pregnancy care; 3. Substance use disorder treatment; 4. Acute psychiatric crisis.

What is the Principle of Double Effect in palliative end-of-life care?

An action with both a good intended effect (e.g., pain relief from high-dose opioids) and an unintended, foreseen potential adverse effect (e.g., respiratory depression shortening life) is ethically permissible provided the intention is solely to relieve suffering.

How should a physician respond when a patient's family requests that a diagnosis of terminal metastatic cancer be withheld from the patient?

Acknowledge the family's concerns, explore their underlying fears and cultural context, but clarify that patients have a fundamental ethical right to know their own medical condition unless the patient explicitly asks not to be told.

What defines a Type I error (alpha) versus a Type II error (beta)?

Type I error (alpha): Rejecting the null hypothesis when it is actually true ('false positive finding'). Type II error (beta): Failing to reject the null hypothesis when it is false ('false negative finding').

What are the four primary ethical principles in modern clinical bioethics?

1. Autonomy: Respect for patients' right to self-determination and decision-making; 2. Beneficence: Acting in the best interest of the patient; 3. Non-maleficence: Do no harm; 4. Justice: Fair and equitable distribution of health resources and burdens.

What are the four essential elements required to establish Decision-Making Capacity in a patient?

Patient must: 1. Communicate a clear and consistent choice; 2. Understand relevant clinical information and diagnoses; 3. Appreciate their specific medical situation and consequences of options; 4. Rationally manipulate options to reach a decision.

What characterizes Phase I, II, III, and IV Clinical Drug Trials?

Phase I: Small group of healthy volunteers; tests safety, pharmacokinetics, and maximum tolerated dose. Phase II: Moderate group of patients with disease; tests efficacy and dosing. Phase III: Large randomized multicenter trial; compares efficacy against standard of care. Phase IV: Post-marketing surveillance; detects rare, long-term adverse effects.

What is the appropriate communication protocol for disclosing a serious medical error to a patient and family?

Disclose the error promptly and honestly, explain clearly what happened in non-technical terms, apologize sincerely for the occurrence, describe the steps taken to mitigate harm and prevent recurrence, and offer appropriate immediate follow-up care.

How are Positive Likelihood Ratio (LR+) and Negative Likelihood Ratio (LR-) calculated from sensitivity and specificity?

LR+ = Sensitivity / (1 - Specificity). LR- = (1 - Sensitivity) / Specificity. LR+ > 10 strongly confirms disease; LR- < 0.1 strongly excludes disease, independent of prevalence.

Can parents legally refuse life-saving medical treatment (such as a blood transfusion) for their minor child based on religious beliefs?

No. The state's parens patriae doctrine empowers physicians and courts to protect children from preventable death or serious harm. In an emergency, life-saving blood must be transfused immediately without awaiting court intervention.

What is Lead-Time Bias and how does it create an illusion of increased survival?

Screening detects a disease earlier in its course without prolonging actual lifespan, making survival time from diagnosis appear longer simply because the diagnosis clock started earlier.

What is the difference between a Funnel Plot and a Forest Plot in meta-analyses?

A Forest plot displays individual study effect sizes and confidence intervals alongside the pooled summary estimate. A Funnel plot evaluates publication bias by plotting study sample size (or standard error) against effect size; asymmetry indicates potential publication bias.

How does Effect Modification differ fundamentally from Confounding?

Effect modification is a genuine biological phenomenon where the association between exposure and outcome varies across strata of a third variable (stratified analysis yields significantly different stratum-specific odds/relative risks); confounding shows identical stratum-specific estimates that differ from the crude estimate.

What is Selection Bias, and what is Berkson's bias?

Selection bias occurs when study subjects are not representative of the target population. Berkson's bias is a specific selection bias where hospitalized patients are used as controls, leading to spuriously inflated association due to higher co-morbidity rates.

What does a 95% Confidence Interval (CI) indicate for an Odds Ratio or Relative Risk, and when is it statistically significant?

It is statistically significant at p < 0.05 if and only if the 95% confidence interval does NOT cross the null value of 1.0 (e.g., 95% CI 1.25–2.40 is significant; 95% CI 0.95–1.80 is not).

What is the order of priority for surrogate decision-makers when an incapacitated patient has no designated healthcare proxy?

Typically: 1. Spouse; 2. Adult children; 3. Parents; 4. Adult siblings; 5. Other nearest adult relatives. The surrogate must apply 'substituted judgment' (what the patient would have wanted), not their own personal preference.

When comparing continuous variables between groups, which statistical tests are used for 2 groups versus 3 or more groups?

For 2 independent groups: Two-sample Student's t-test (or Mann-Whitney U test if non-parametric). For 3 or more independent groups: One-way Analysis of Variance (ANOVA; or Kruskal-Wallis if non-parametric). Chi-square test is used for categorical variables.

What is Statistical Power and how is it related to Type II error?

Power = 1 - beta. The probability of correctly rejecting a false null hypothesis (detecting a true difference when one exists). Increased by larger sample size, larger effect size, and higher significance level (alpha).

What is the Hawthorne Effect?

Observer bias where study participants alter their behavior simply because they know they are being monitored or evaluated.

What is Intention-to-Treat (ITT) analysis and what is its primary benefit?

All participants are analyzed in the groups to which they were originally randomized, regardless of whether they completed treatment or dropped out. Preserves randomization benefits and avoids attrition bias, reflecting real-world clinical effectiveness.

USMLE Step 2 CK — Biostatistics, Epidemiology, Study Design & Medical Ethics

Sensitivity, specificity, predictive values, likelihood ratios, odds/hazard ratios, study designs, bias types, phase trials, autonomy, and medical ethics dilemmas.

Level: advanced
Languages: en → en
Cards: 32
Markdown: USMLE Step 2 CK — Biostatistics, Epidemiology, Study Design & Medical Ethics (.md)

Flashcards (32)

#	Front (Term / Question)	Back (Definition / Answer)
1	How is Relative Risk (RR) calculated from a Cohort study?	RR = Risk in exposed / Risk in unexposed = [a / (a + b)] / [c / (c + d)]. If RR = 1, risk is identical; RR > 1 indicates increased risk; RR < 1 indicates protective association.
2	How is Absolute Risk Reduction (ARR) and Number Needed to Treat (NNT) calculated?	ARR = Event rate in control group (CER) - Event rate in treatment group (EER). NNT = 1 / ARR (always rounded UP to the nearest whole integer).
3	How is the Odds Ratio (OR) calculated from a 2x2 table in a Case-Control study?	OR = (a * d) / (b * c), where 'a' = exposed cases, 'b' = exposed controls, 'c' = unexposed cases, and 'd' = unexposed controls. Approximates Relative Risk when disease prevalence is low ('rare disease assumption').
4	Can an adult Jehovah's Witness patient with full decision-making capacity refuse life-saving blood transfusion?	Yes. Autonomy empowers a competent adult patient to refuse any medical treatment, even if that refusal results in inevitable death. The physician must verify capacity and explore acceptable alternatives (cell saver, erythropoietin), but cannot override their refusal.
5	Under what circumstances may a physician breach patient confidentiality without patient consent?	1. Reportable infectious diseases (e.g., tuberculosis, syphilis, measles); 2. Suspected child abuse or elder abuse (mandatory reporting); 3. Clear, imminent, serious threat of self-harm or violence against an identifiable third party ('Tarasoff duty to warn'); 4. Impaired drivers unfit to operate vehicles.
6	How are Sensitivity and Specificity defined and calculated from a 2x2 contingency table?	Sensitivity = TP / (TP + FN); the proportion of true disease cases correctly detected by the test. Specificity = TN / (TN + FP); the proportion of non-diseased individuals correctly identified as negative.
7	How does Decision-Making Capacity differ from Competence?	Capacity is a clinical determination made by a healthcare provider for a specific medical decision at a specific point in time. Competence is a legal determination made by a judge in a court of law.
8	What is Confounding and how can it be controlled during study design and statistical analysis?	A third variable associated with both exposure and outcome, but not on the causal pathway. Controlled in design by Randomization, Matching, and Restriction; controlled in analysis by Stratification and Multivariable Regression.
9	How does changing disease prevalence affect Sensitivity, Specificity, Positive Predictive Value (PPV), and Negative Predictive Value (NPV)?	Sensitivity and Specificity are intrinsic test properties and do NOT change with prevalence. Higher prevalence increases PPV and decreases NPV; lower prevalence decreases PPV and increases NPV.
10	What is Length-Time Bias in cancer screening programs?	Screening tests disproportionately detect slowly progressing, indolent cases with long asymptomatic phases, while missing rapidly progressive aggressive cancers that present symptomatically between screening intervals.
11	What is Recall Bias and in which study design is it most commonly encountered?	A form of measurement/information bias where cases recall past exposures differently (often more thoroughly or falsely) than healthy controls; classically plagues retrospective Case-Control studies.
12	How is Number Needed to Harm (NNH) calculated?	NNH = 1 / Attributable Risk (where Attributable Risk = Adverse event rate in exposed - Adverse event rate in unexposed). Always rounded DOWN to the nearest integer for conservative harm estimation.
13	Under what conditions can a minor receive medical treatment without parental knowledge or consent ('emancipated minor' vs. specific exceptions)?	Emancipation (married, self-supporting/military, or court decree) grants full consent rights. Regardless of emancipation, most jurisdictions allow confidential minor consent for: 1. Sexually transmitted infection treatment; 2. Contraception/pregnancy care; 3. Substance use disorder treatment; 4. Acute psychiatric crisis.
14	What is the Principle of Double Effect in palliative end-of-life care?	An action with both a good intended effect (e.g., pain relief from high-dose opioids) and an unintended, foreseen potential adverse effect (e.g., respiratory depression shortening life) is ethically permissible provided the intention is solely to relieve suffering.
15	How should a physician respond when a patient's family requests that a diagnosis of terminal metastatic cancer be withheld from the patient?	Acknowledge the family's concerns, explore their underlying fears and cultural context, but clarify that patients have a fundamental ethical right to know their own medical condition unless the patient explicitly asks not to be told.
16	What defines a Type I error (alpha) versus a Type II error (beta)?	Type I error (alpha): Rejecting the null hypothesis when it is actually true ('false positive finding'). Type II error (beta): Failing to reject the null hypothesis when it is false ('false negative finding').
17	What are the four primary ethical principles in modern clinical bioethics?	1. Autonomy: Respect for patients' right to self-determination and decision-making; 2. Beneficence: Acting in the best interest of the patient; 3. Non-maleficence: Do no harm; 4. Justice: Fair and equitable distribution of health resources and burdens.
18	What are the four essential elements required to establish Decision-Making Capacity in a patient?	Patient must: 1. Communicate a clear and consistent choice; 2. Understand relevant clinical information and diagnoses; 3. Appreciate their specific medical situation and consequences of options; 4. Rationally manipulate options to reach a decision.
19	What characterizes Phase I, II, III, and IV Clinical Drug Trials?	Phase I: Small group of healthy volunteers; tests safety, pharmacokinetics, and maximum tolerated dose. Phase II: Moderate group of patients with disease; tests efficacy and dosing. Phase III: Large randomized multicenter trial; compares efficacy against standard of care. Phase IV: Post-marketing surveillance; detects rare, long-term adverse effects.
20	What is the appropriate communication protocol for disclosing a serious medical error to a patient and family?	Disclose the error promptly and honestly, explain clearly what happened in non-technical terms, apologize sincerely for the occurrence, describe the steps taken to mitigate harm and prevent recurrence, and offer appropriate immediate follow-up care.
21	How are Positive Likelihood Ratio (LR+) and Negative Likelihood Ratio (LR-) calculated from sensitivity and specificity?	LR+ = Sensitivity / (1 - Specificity). LR- = (1 - Sensitivity) / Specificity. LR+ > 10 strongly confirms disease; LR- < 0.1 strongly excludes disease, independent of prevalence.
22	Can parents legally refuse life-saving medical treatment (such as a blood transfusion) for their minor child based on religious beliefs?	No. The state's parens patriae doctrine empowers physicians and courts to protect children from preventable death or serious harm. In an emergency, life-saving blood must be transfused immediately without awaiting court intervention.
23	What is Lead-Time Bias and how does it create an illusion of increased survival?	Screening detects a disease earlier in its course without prolonging actual lifespan, making survival time from diagnosis appear longer simply because the diagnosis clock started earlier.
24	What is the difference between a Funnel Plot and a Forest Plot in meta-analyses?	A Forest plot displays individual study effect sizes and confidence intervals alongside the pooled summary estimate. A Funnel plot evaluates publication bias by plotting study sample size (or standard error) against effect size; asymmetry indicates potential publication bias.
25	How does Effect Modification differ fundamentally from Confounding?	Effect modification is a genuine biological phenomenon where the association between exposure and outcome varies across strata of a third variable (stratified analysis yields significantly different stratum-specific odds/relative risks); confounding shows identical stratum-specific estimates that differ from the crude estimate.
26	What is Selection Bias, and what is Berkson's bias?	Selection bias occurs when study subjects are not representative of the target population. Berkson's bias is a specific selection bias where hospitalized patients are used as controls, leading to spuriously inflated association due to higher co-morbidity rates.
27	What does a 95% Confidence Interval (CI) indicate for an Odds Ratio or Relative Risk, and when is it statistically significant?	It is statistically significant at p < 0.05 if and only if the 95% confidence interval does NOT cross the null value of 1.0 (e.g., 95% CI 1.25–2.40 is significant; 95% CI 0.95–1.80 is not).
28	What is the order of priority for surrogate decision-makers when an incapacitated patient has no designated healthcare proxy?	Typically: 1. Spouse; 2. Adult children; 3. Parents; 4. Adult siblings; 5. Other nearest adult relatives. The surrogate must apply 'substituted judgment' (what the patient would have wanted), not their own personal preference.
29	When comparing continuous variables between groups, which statistical tests are used for 2 groups versus 3 or more groups?	For 2 independent groups: Two-sample Student's t-test (or Mann-Whitney U test if non-parametric). For 3 or more independent groups: One-way Analysis of Variance (ANOVA; or Kruskal-Wallis if non-parametric). Chi-square test is used for categorical variables.
30	What is Statistical Power and how is it related to Type II error?	Power = 1 - beta. The probability of correctly rejecting a false null hypothesis (detecting a true difference when one exists). Increased by larger sample size, larger effect size, and higher significance level (alpha).
31	What is the Hawthorne Effect?	Observer bias where study participants alter their behavior simply because they know they are being monitored or evaluated.
32	What is Intention-to-Treat (ITT) analysis and what is its primary benefit?	All participants are analyzed in the groups to which they were originally randomized, regardless of whether they completed treatment or dropped out. Preserves randomization benefits and avoids attrition bias, reflecting real-world clinical effectiveness.