USMLE Step 1 biostatistics, epidemiology & ethics questions

Randomization allocates participants to study groups by chance, which tends to balance both measured and unmeasured confounding variables between groups, including factors investigators have not thought to measure. This is distinct from matching, which only balances known, pre-specified variables. Blinding, not randomization, addresses observer and assessment bias. Adequate sample size is determined by power calculations, not randomization itself.

• B. Randomization does not affect the alpha level or probability of a type I error
• C. This describes blinding, a separate design feature from randomization
• D. A control group is still required; randomization only determines how subjects are allocated to groups
• E. Sample size adequacy depends on power calculations, not on the randomization process

Relative risk is calculated as the incidence of disease in the exposed group divided by the incidence in the unexposed group. Incidence in smokers = 80/1,000 = 0.08 (8%); incidence in nonsmokers = 20/2,000 = 0.01 (1%). Relative risk = 0.08/0.01 = 8.0, meaning smokers have 8 times the risk of developing lung cancer compared with nonsmokers. Relative risk requires a cohort design because the incidence of disease must be known in both exposed and unexposed groups.

• A. This value does not correspond to any correct ratio of the given incidence proportions
• B. This results from an arithmetic error in calculating the incidence proportions
• C. This results from incorrectly dividing the number of cases (80/20) rather than the incidence proportions
• D. This is the reciprocal of the correct relative risk, representing the risk in nonsmokers relative to smokers

In a case-control study, the odds ratio is calculated as (a × d)/(b × c), where a = exposed cases, b = unexposed cases, c = exposed controls, and d = unexposed controls. Here, (60 × 80)/(40 × 20) = 4,800/800 = 6.0. This means the odds of raw shellfish consumption are 6 times higher in people with hepatitis A than in those without, supporting an association. Case-control studies cannot directly calculate relative risk because incidence data are not collected, so the odds ratio is used instead.

• A. This results from an incorrect cross-multiplication of the 2x2 table values
• B. This results from doubling the correct odds ratio due to a calculation error
• C. This is the reciprocal of the correct odds ratio
• D. This results from comparing exposure proportions directly rather than computing the cross-product odds ratio

A cross-sectional study collects data on exposure and disease status at a single point in time from a defined population, making it ideal for calculating prevalence, the proportion of a population affected by a condition at that moment. Because there is no follow-up period, cross-sectional studies cannot measure incidence, the rate of new cases arising over time, nor can they calculate relative risk or attributable risk, which require longitudinal data. This design is useful for generating hypotheses about associations but is limited by its inability to establish temporal sequence between exposure and outcome.

• A. Incidence requires following a population over time to identify new cases, which a cross-sectional study does not do
• B. Relative risk requires incidence data from exposed and unexposed cohorts followed longitudinally
• D. Cumulative incidence requires a defined follow-up period, which is absent in a cross-sectional design
• E. Attributable risk requires incidence data, which is not available from a single-time-point survey

Late-look bias occurs when a study samples patients at a point in time that systematically excludes those who died early from rapidly fatal disease, leaving a sample enriched for individuals with a longer, more indolent course. Here, patients who died quickly after symptom onset (likely those with the most aggressive tumors) never reached the oncology clinic and were excluded, biasing the study toward patients with longer survival and more time between symptom onset and diagnosis. This differs from length-time bias, which specifically refers to screening programs preferentially detecting slow-growing tumors. The researcher's conclusion reflects the survivors remaining in the sample, not the true natural history of the disease.

• A. Recall bias involves differential memory accuracy between groups, not exclusion of a subgroup from the sample.
• B. Confounding by indication concerns treatment-outcome associations distorted by why treatment was chosen, which is not the issue described.
• D. Lead-time bias applies to screening-detected cases appearing to survive longer because they are diagnosed earlier in their course, not to exclusion of patients who died before diagnosis.
• E. Length-time bias is a screening-specific phenomenon where slow-growing tumors are preferentially caught; this study is not a screening program.

Specificity is the proportion of patients without disease who correctly test negative, calculated as true negatives divided by all patients without disease (true negatives plus false positives). Here, among the 350 patients without myocardial infarction, 70 tested positive (false positives), so 280 tested negative (true negatives). Specificity = 280 / 350 = 80.0%. Sensitivity, in contrast, is 135/150 = 90.0%, which is the proportion of diseased patients correctly identified as positive. Confusing the denominator (using all patients rather than only those without disease) is a common calculation error for specificity.

• A. This value equals the sensitivity (135/150), not the specificity.
• B. This value results from an incorrect denominator (e.g., using true negatives over total positives) rather than the diseased-free population.
• C. This value does not match the correct calculation of true negatives over total disease-free patients.
• E. This value does not correspond to true negatives divided by total disease-free patients.

Positive predictive value is directly influenced by disease prevalence in the tested population, even when sensitivity and specificity remain fixed properties of the test. At higher prevalence (40% in intravenous drug users), a larger proportion of positive results are true positives because there are proportionally more diseased individuals to generate true positives relative to the fixed false-positive rate from the smaller disease-free group. At lower prevalence (2% in blood donors), the same specificity yields a relatively larger pool of false positives among the much larger disease-free population, driving PPV down. This is why PPV rises with prevalence and falls with declining prevalence, unlike sensitivity and specificity which are prevalence-independent.

• A. Absolute false-positive counts depend on sample size, not on the proportional relationship that determines PPV; the directionality stated is also incorrect.
• B. PPV can be calculated from prevalence, sensitivity, and specificity alone without needing the raw sample size.
• D. Sensitivity and specificity are intrinsic to the test, but PPV and NPV depend on prevalence and therefore differ between populations.
• E. Specificity being prevalence-independent does not mean PPV is prevalence-independent; PPV depends directly on prevalence.

Positive predictive value is calculated as true positives divided by all patients who test positive (true positives plus false positives). Here, true positives = 92 and false positives = 54, so total positive tests = 146. PPV = 92 / 146 ≈ 63.0%. Negative predictive value, by comparison, uses true negatives (900 − 54 = 846) divided by all negative tests (846 + 8 false negatives = 854), giving NPV ≈ 99.1%. PPV reflects the probability that a patient with a positive test result truly has the disease.

• A. This value does not correspond to the correct PPV calculation using the given true and false positive counts.
• B. This value does not match true positives divided by all positive test results.
• C. This value does not correspond to the correct ratio of true positives to total positive tests.
• D. This value approximates sensitivity (92/100), not positive predictive value.

The likelihood ratio approach converts pretest probability to odds, multiplies by the likelihood ratio, and converts back to probability. Pretest probability of 30% corresponds to pretest odds of 0.30/0.70 ≈ 0.43. Posttest odds = 0.43 × 6.0 ≈ 2.57. Converting back to probability: posttest probability = odds / (1 + odds) = 2.57 / 3.57 ≈ 72%. A likelihood ratio greater than 1 increases the probability of disease after a positive test, and larger likelihood ratios produce larger shifts in probability, which is why LRs above 10 are considered strong rule-in evidence.

• A. This value underestimates the increase in probability produced by a likelihood ratio of 6.0.
• B. This value simply restates the pretest probability and ignores the effect of the positive test result.
• D. This value overestimates the posttest probability given the calculated posttest odds of approximately 2.57.
• E. This value substantially overestimates the posttest probability for this likelihood ratio and pretest probability.

When a missed diagnosis carries severe consequences, as with pulmonary embolism, the priority is a test with high sensitivity, since a highly sensitive test has few false negatives and a negative result can be used to reliably exclude disease (a negative result on a highly sensitive test rules disease out, following the mnemonic SnNout). D-dimer testing is a classic example used for this purpose in low-to-moderate pretest probability patients. Specificity and positive predictive value are more relevant when the goal is confirming disease, which matters when false positives would be costly, but that is not the priority in an urgent rule-out scenario. Negative predictive value is prevalence-dependent and therefore not an intrinsic test characteristic that guides the initial choice of test.

• A. High specificity minimizes false positives, which is prioritized when confirming disease, not ruling it out.
• B. A high positive likelihood ratio is relevant to confirming disease after a positive test, not to safely excluding disease.
• D. High PPV is relevant to confirming disease after a positive result, not excluding disease when a negative result is the goal.
• E. Negative predictive value depends on prevalence and is not the fixed test property used to select a screening test for ruling out disease.

Screening tests such as ELISA are designed with high sensitivity to minimize false negatives, which comes at the cost of lower specificity and a higher false-positive rate, especially in low-prevalence populations such as an asymptomatic patient without risk factors. A confirmatory test with high specificity, such as an HIV antibody differentiation assay (or historically, western blot), is then used to minimize false positives and verify that a positive screening result reflects true disease (a positive result on a highly specific test confirms disease, following the mnemonic SpPin). This two-step strategy, highly sensitive screening followed by highly specific confirmation, is a standard approach in diagnostic testing for low-prevalence conditions. Acting on the screening result alone risks a false diagnosis with major psychological and social consequences for the patient.

• A. Repeating the identical screening assay does not address its lower specificity and would not resolve the risk of a false positive.
• C. A single positive result on a screening test, particularly in a low-prevalence patient, is not sufficient to confirm the diagnosis without a confirmatory test.
• D. CD4 count assesses immune status in confirmed HIV infection but does not establish the diagnosis itself.
• E. Starting antiretroviral therapy before confirming the diagnosis is inappropriate and exposes the patient to unnecessary risk if the screening result is a false positive.

The area under the receiver operating characteristic (ROC) curve (AUC) is a summary measure of a test's overall discriminatory ability across all possible cutoff thresholds, with 1.0 representing a perfect test and 0.5 representing no discriminative value (equivalent to chance). An AUC of 0.91 indicates substantially better overall accuracy than an AUC of 0.68. A curve that lies above and to the left of another reflects a better trade-off between sensitivity and specificity at every cutoff. AUC does not by itself determine predictive values, which depend on disease prevalence, nor does it specify a particular optimal cutoff.

• A. AUC does not indicate the specific cutoff value needed for maximum sensitivity.
• B. AUC reflects validity (accuracy), not reliability (reproducibility).
• D. A higher AUC reflects better overall performance, which corresponds to a higher, not lower, likelihood ratio for a positive test at comparable cutoffs.
• E. Positive predictive value depends on disease prevalence, not solely on AUC.

Lowering the cutoff value for a positive test result makes it easier for individuals to be classified as test-positive, which increases the number of true positives captured (raising sensitivity) but also increases the number of false positives (lowering specificity). This inverse relationship between sensitivity and specificity is intrinsic to any continuous diagnostic test and is graphically represented by moving along a single ROC curve rather than switching to a different curve. The choice of cutoff reflects a trade-off tailored to the clinical goal, such as maximizing sensitivity for a screening test used to rule out disease.

• A. Lowering the cutoff increases, not decreases, sensitivity.
• B. Sensitivity and specificity move in opposite directions when the cutoff changes; both cannot increase simultaneously for the same test.
• C. Lowering the cutoff increases sensitivity, not decreases it.
• E. Changing the cutoff alters the test's sensitivity and specificity; prevalence is a property of the population, not the test.

Attributable risk (risk difference) is calculated as the incidence of disease in the exposed group minus the incidence in the unexposed group: 8 per 1,000 minus 2 per 1,000 equals 6 per 1,000 women. This measure estimates the excess disease incidence that can be directly attributed to the exposure, assuming the association is causal. It is useful for quantifying the absolute impact of removing an exposure from a population, distinguishing it from relative measures such as relative risk.

• A. This value does not result from the correct risk difference calculation.
• B. This value does not correspond to the correct subtraction of incidence rates (8 - 2 = 6, not 4).
• C. This describes the relative risk (8/2 = 4), not the attributable risk, which is an absolute measure.
• D. This represents the sum, not the difference, of the two incidence rates.

The number needed to treat (NNT) is calculated as the reciprocal of the absolute risk reduction (ARR). Here, ARR = 15% - 9% = 6%, or 0.06. NNT = 1/0.06 ≈ 16.7, which rounds to 17 patients. This means that approximately 17 patients must receive the new drug for 3 years to prevent one additional stroke compared with placebo, providing a clinically intuitive measure of treatment benefit at the population level.

• A. This value appears to come from using relative risk reduction instead of absolute risk reduction.
• B. This value does not correspond to the reciprocal of the correct absolute risk reduction.
• D. This value does not correspond to the reciprocal of the correct absolute risk reduction.
• E. This value does not correspond to the correct calculation of 1/ARR.

The number needed to harm (NNH) is the reciprocal of the absolute risk increase (ARI). Here, ARI = 12% - 4% = 8%, or 0.08. NNH = 1/0.08 = 12.5, which rounds to 13 patients. This value indicates that for every 13 patients treated with the NSAID instead of acetaminophen for 1 year, one additional case of gastrointestinal bleeding would be expected to occur, providing a practical measure of a treatment's harm burden.

• A. This value does not correspond to the correct calculation of 1/ARI.
• C. This value does not correspond to the reciprocal of the correct absolute risk increase.
• D. This value does not correspond to the correct calculation of 1/ARI.
• E. This value represents the relative risk (12/4 = 3), not the number needed to harm.

Prevalence approximates incidence multiplied by average disease duration when the population is stable. Because disease B persists for decades while disease A resolves within days, disease B accumulates many more existing cases at any given point in time despite having the same incidence rate. This illustrates why prevalence is influenced not only by how many new cases occur but also by how long individuals remain affected, which is a key distinction between incidence (new cases over time) and prevalence (existing cases at a point in time).

• A. Case-fatality rate affects disease duration but is not required to compare the relative prevalence of the two diseases qualitatively.
• B. With constant incidence and chronic duration, prevalence of disease B would tend to rise toward a steady state, not decrease.
• C. Disease A has a much shorter duration, so despite equal incidence, its prevalence will be lower, not higher, than disease B's.
• E. Equal incidence does not translate into equal prevalence when disease durations differ markedly.

A p-value of 0.001 that is less than the prespecified alpha of 0.05 means the observed difference between groups is statistically significant, i.e., unlikely to have arisen from chance variation alone if the null hypothesis were true. The p-value is the probability of obtaining a result as extreme as, or more extreme than, the one observed, assuming the null hypothesis is true; it is not the probability that the null or alternative hypothesis is true. Statistical significance also does not automatically equate to clinical significance, which depends on the magnitude of effect and its practical importance.

• A. The p-value does not give the probability that the alternative hypothesis is true.
• B. Statistical significance does not necessarily indicate clinical importance; that requires judgment about the magnitude of effect.
• C. The p-value is not the probability that the null hypothesis is true; this is a common misinterpretation.
• D. A small p-value reflects the strength of evidence against the null hypothesis in this sample, not reproducibility of results.

A type I error occurs when a true null hypothesis is incorrectly rejected, concluding that an association exists when in reality it does not. The probability of making a type I error is denoted by alpha, the significance level set before the study (commonly 0.05). In this scenario, the null hypothesis (no true association) is true, but the investigators rejected it based on a statistically significant p-value, which is the definition of a type I error, often called a false-positive result.

• A. A type II error is failing to reject a false null hypothesis, which is the opposite of what occurred here.
• C. Regression to the mean describes extreme values moving toward the average on repeat measurement, unrelated to this scenario.
• D. Low power relates to the probability of detecting a true effect and would increase type II error, not type I error.
• E. Confounding bias refers to a third variable distorting the true association, which is not described here.

A type II error occurs when a study fails to reject a false null hypothesis, concluding there is no effect when a true effect actually exists. This is often called a false-negative result and its probability is denoted beta. Small sample sizes, as in this 20-patient pilot study, reduce statistical power (1 - beta) and increase the likelihood of a type II error because the study lacks sufficient ability to detect a true difference.

• A. Recall bias involves differential memory of past exposures, not applicable to this randomized comparison.
• B. Selection bias relates to how participants are chosen for a study, not to a statistical conclusion about effect detection.
• C. A type I error is incorrectly rejecting a true null hypothesis, which is the opposite of what happened here.
• D. Lead-time bias relates to earlier diagnosis appearing to improve survival in screening studies, unrelated to this scenario.

Increasing the sample size reduces the standard error of the estimated effect, which narrows the confidence interval and increases the ability of the study to detect a true difference between groups if one exists. This increase in statistical power (1 - beta) directly corresponds to a decrease in the probability of a type II error (beta), since power and beta are complementary. Sample size does not directly change the type I error rate (alpha), which is set by the investigator as the significance threshold.

• A. Power and type II error move in opposite directions, so both cannot increase together.
• B. Increasing sample size increases power rather than decreasing it, and alpha is unaffected by sample size.
• C. Power depends on sample size, effect size, variability, and alpha; sample size is a major determinant of power.
• E. Alpha, the type I error rate, is set by the investigator and is not directly changed by increasing sample size.

Because the 95% confidence interval for the odds ratio (1.8 to 9.6) does not include the null value of 1, the association between asbestos exposure and mesothelioma is statistically significant at the conventional alpha of 0.05. For ratio measures of association such as odds ratios and relative risks, a value of 1 indicates no association; if the entire confidence interval lies above or below 1, the result is unlikely to be due to chance alone. The width of the interval also reflects the precision of the estimate, with narrower intervals indicating more precise estimates.

• B. There is no information given about confounding, and the interval itself does not indicate directionality of bias.
• C. Because the interval excludes 1, the association is in fact statistically significant, not non-significant.
• D. A confidence interval provides a range of plausible values for the true odds ratio, not certainty about one exact value.
• E. An interval excluding the null value with a wide range reflects significance, not necessarily low power; low power typically produces wide intervals that include the null.

The Student's t-test is used to compare the means of a continuous, normally distributed variable between exactly two independent groups, which matches this scenario of comparing mean hemoglobin A1c between two treatment groups. The chi-square test is reserved for categorical data, ANOVA is used when comparing means among three or more groups, Pearson correlation measures the linear relationship between two continuous variables rather than comparing group means, and Kaplan-Meier analysis estimates survival probabilities over time.

• B. Kaplan-Meier analysis is used to estimate time-to-event probabilities, not to compare mean laboratory values between two groups.
• C. Chi-square compares proportions of categorical variables between groups, not means of continuous variables.
• D. Pearson correlation quantifies the strength and direction of a linear association between two continuous variables, not a difference in means between groups.
• E. ANOVA is used when comparing means among three or more groups, not just two.

Analysis of variance (ANOVA) is the appropriate test for comparing means of a continuous, normally distributed variable among three or more independent groups, as in this trial comparing fasting glucose across four dose groups. ANOVA tests whether at least one group mean differs from the others by comparing between-group variance to within-group variance, generating a single overall p-value rather than requiring multiple pairwise t-tests, which would otherwise inflate the type I error rate. If the ANOVA result is significant, post hoc tests can then identify which specific group means differ.

• A. The Wilcoxon rank-sum test is a nonparametric test comparing two groups, not three or more, and is used when data are not normally distributed.
• B. Chi-square test compares categorical proportions, not means of continuous variables.
• C. Pearson correlation quantifies association between two continuous variables, not differences in means among groups.
• D. A Student's t-test compares means between only two groups and would require multiple comparisons here, inflating the type I error rate.

The Pearson correlation coefficient (r) measures the strength and direction of a linear relationship between two continuous variables, ranging from -1 to +1. An r value of 0.82 indicates a strong positive association, meaning both variables increase together. Values closer to +1 or -1 indicate stronger linear relationships, while values near 0 indicate little to no linear relationship. Correlation does not establish causation, as confounding variables or reverse causation may explain the association.

• A. A negative r value would indicate an inverse relationship, where one variable increases as the other decreases; here r is positive.
• B. Correlation coefficients describe association, not causation; a controlled experiment would be needed to establish causality.
• D. An r of 0.82 is close to 1, indicating a strong, not weak, correlation.
• E. An r value of 0.82 indicates a strong relationship, not an absence of correlation.

In a forest plot, the diamond represents the pooled effect estimate, and its position relative to the line of no effect (RR = 1) indicates statistical significance; here, the diamond and its 95% CI lie entirely below 1, indicating a significant reduction in stroke risk. The degree of overlap among individual trial confidence intervals reflects heterogeneity: substantial overlap with variability close to what is expected by chance indicates low heterogeneity (low I² statistic), supporting that pooling the results is statistically appropriate. High heterogeneity would appear as widely scattered point estimates with little overlap, often prompting use of a random-effects model instead of a fixed-effect model.

• A. The substantial overlap among individual trial estimates indicates low, not high, heterogeneity.
• B. The pooled confidence interval excludes 1, indicating a statistically significant effect, not a null result.
• C. The relative risk is less than 1, indicating a reduction, not an increase, in stroke risk.
• D. The low heterogeneity and significant pooled estimate support a conclusive, not inconclusive, result.

The mean is calculated by summing all values and dividing by the number of observations, so an extreme outlier like 40 days disproportionately pulls the mean upward. The median, defined as the middle value when data are ordered, is resistant to extreme values because it depends only on rank order, not magnitude; here the median remains 5 days regardless of how extreme the ninth value is. This property makes the median a more robust measure of central tendency for skewed data sets, such as hospital length-of-stay data, which commonly has a long right tail.

• A. The mean is affected by outliers; it increases substantially due to the extreme value.
• B. The median is unaffected by the magnitude of the outlier, only its rank; it does not increase markedly.
• C. Both values should increase or stay stable with a high-value outlier; the mean does not decrease.
• E. The median is resistant to outliers and does not change markedly with a single extreme value.

In a positively (right) skewed distribution, the tail extends toward higher values, pulling the mean above the median, as seen here with a mean of $78,000 exceeding the median of $58,000. This pattern is typical of income data, where most individuals earn modest amounts but a small number of high earners create a long right tail. In contrast, a negatively (left) skewed distribution has a tail extending toward lower values, causing the mean to fall below the median. Recognizing skew direction is important because it affects which measure of central tendency (mean vs median) best represents the typical value.

• A. A uniform distribution has roughly equal frequency across all values, which is inconsistent with a long tail and unequal mean/median.
• B. A left-skewed distribution would show the mean lower than the median, not higher.
• C. A normal distribution is symmetric, with the mean equal to the median, which is not the case here.
• E. A bimodal distribution has two distinct peaks, which is not described here.

The standard error of the mean (SEM) is calculated as the standard deviation divided by the square root of the sample size (SEM = SD/√n). Increasing the sample size from 25 to 400 represents a 16-fold increase in n, and since √16 = 4, the SEM decreases to one-fourth of its original value, assuming the standard deviation remains stable. Unlike the standard deviation, which reflects the inherent variability of individual measurements in the population and does not systematically change with sample size, the SEM reflects the precision of the sample mean as an estimate of the true population mean and improves (decreases) as sample size increases.

• A. The square root of the 16-fold increase in sample size is 4, so the SEM decreases to one-fourth, not one-half, of its original value.
• B. The SEM decreases, not increases, as sample size increases.
• C. The SEM decreases with increasing sample size regardless of whether the standard deviation changes, as long as the standard deviation remains stable, which is the case described.
• E. The SEM decreases substantially with a 16-fold increase in sample size; it does not remain unchanged.

Phase I trials represent the first stage of human testing for a new drug and are conducted in a small number of healthy volunteers (typically 20-100) to determine safety, tolerability, pharmacokinetics, and an appropriate dosage range. These trials use escalating doses to identify the maximum tolerated dose and to detect dose-limiting toxicities before the drug is tested in patients with the target disease. Only after a drug demonstrates acceptable safety in Phase I does testing proceed to Phase II trials, which involve patients with the disease to assess efficacy.

• A. Phase III trials involve large numbers of patients to compare the new treatment with standard therapy, not healthy volunteers.
• C. Phase II trials involve patients with the target disease to assess efficacy, not healthy volunteers assessing safety and dosage.
• D. Phase IV trials occur after FDA approval to monitor long-term safety and effectiveness in the general population.
• E. Preclinical trials involve laboratory and animal studies, not human testing, and were already completed in this scenario.

Phase II trials are conducted after Phase I establishes basic safety and involve a relatively small group of patients who actually have the target disease, in this case rheumatoid arthritis. The primary goals of Phase II are to evaluate the drug's efficacy for the intended indication and to further assess short-term side effects in the disease population, often using dose-ranging designs to identify the optimal therapeutic dose. If a drug demonstrates efficacy and an acceptable safety profile in Phase II, it then proceeds to Phase III, which involves larger, randomized trials comparing it with standard treatment.

• A. Preclinical trials involve laboratory and animal studies, not testing in human patients.
• C. Phase IV trials occur after FDA approval to monitor long-term outcomes in the general population, not during initial efficacy testing.
• D. Phase I trials involve healthy volunteers to assess safety and dosage, not patients with the disease assessing efficacy.
• E. Phase III trials involve much larger patient populations and are designed to confirm efficacy and monitor adverse effects on a broader scale, typically comparing against standard therapy.

Physicians are legally mandated to report suspected child abuse to child protective services or law enforcement whenever there is reasonable suspicion, regardless of parental consent or objection. The physical findings here, bruises of varying ages, a spiral fracture inconsistent with the stated mechanism, and behavioral cues, raise reasonable suspicion. Mandatory reporting laws override the usual requirement for parental consent and confidentiality protections because protecting the child from ongoing harm takes priority. The reporting threshold is reasonable suspicion, not certainty.

• A. Discharging the child without reporting leaves her at risk for further abuse and violates mandatory reporting law.
• B. Parental consent is not required or legally relevant for mandatory abuse reporting.
• C. Reporting does not require permission from either parent; the duty is to the state, not the family.
• D. Documentation alone is insufficient; a report must be filed promptly given reasonable suspicion of abuse.

Active tuberculosis is a legally reportable communicable disease in all U.S. states, and physicians must notify public health authorities regardless of patient objection. This exception to confidentiality exists because uncontrolled transmission poses a direct risk to the community, and public health agencies need this information to perform contact tracing, ensure treatment adherence (including directly observed therapy if needed), and prevent outbreaks. Patient autonomy and confidentiality are outweighed by the duty to protect public health in the case of reportable infectious diseases.

• A. Withholding a mandatory report violates public health law and places others at risk of infection.
• B. No court order is needed; mandatory disease reporting is a standing legal requirement.
• C. Reporting is required at diagnosis, not contingent on treatment failure.
• E. Reporting decisions are a legal duty of the physician and health department, not the employer.

Physicians have an ethical and often legal obligation to report colleagues suspected of practicing while impaired, because impaired practice endangers patient safety. The appropriate channel is typically a hospital's physician health program, credentialing committee, or state medical board, which can arrange evaluation and treatment while protecting patients. This duty exists independently of whether harm has already occurred; the goal is prevention. Reporting is considered an act of professional responsibility, not betrayal, and is distinct from simply confronting the colleague informally, which does not ensure patient safety or get the colleague appropriate help.

• A. Waiting for harm to occur before acting fails the duty to prevent foreseeable patient injury.
• B. The reporting duty belongs to the physician who witnessed the impairment, not to nursing staff alone.
• D. An informal denial does not resolve the safety concern or ensure appropriate evaluation and treatment.
• E. Professional courtesy does not override the ethical duty to protect patients from impaired practice.

Current professional guidelines, including those from the American Medical Association, require full disclosure of medical errors to patients regardless of whether harm resulted. Honest disclosure respects patient autonomy, maintains trust, and allows the patient to make informed decisions about their ongoing care, even when the clinical outcome is favorable. The absence of lasting harm does not eliminate the ethical obligation to be truthful; withholding information because an error was minor is a form of deception that undermines the patient's right to know what happened during their own care.

• A. Documentation alone does not fulfill the ethical obligation to communicate honestly with the patient.
• C. Lack of lasting harm does not remove the ethical duty to disclose; nondisclosure is deceptive regardless of outcome.
• D. Internal quality reporting does not substitute for the physician's direct duty to inform the patient.
• E. Disclosure should be proactive, not contingent on the patient happening to ask the right question.

Modest, inexpensive gifts that express gratitude, such as a small box of cookies, are generally acceptable in the physician-patient relationship and do not compromise professional judgment. Large monetary gifts, however, raise concern that they could influence clinical decision-making, create a sense of obligation, or be perceived as paying for preferential treatment, and should be declined. The physician should explain that the quality and attention of care are not contingent on gifts, which preserves trust and avoids any appearance of a conflict of interest.

• B. Redirecting the cash to the hospital account does not resolve the ethical problem of accepting a disproportionate gift tied to care expectations.
• C. Accepting a large cash gift risks the appearance that care could be influenced by payment, which is inappropriate.
• D. Family approval is irrelevant to whether accepting a large monetary gift from a patient is ethically appropriate.
• E. Declining a token gift like cookies is unnecessarily rigid and can damage rapport without ethical justification.

The clinical picture, clear rhinorrhea, low-grade fever, absence of exudate or lymphadenopathy, is consistent with a viral upper respiratory infection, for which antibiotics provide no benefit and carry risks of adverse effects and antimicrobial resistance. The appropriate approach is shared decision-making: educate the patient on the viral etiology and natural course, address her underlying concern (missing her trip), and offer symptomatic treatment. This respects her autonomy while avoiding inappropriate prescribing driven solely by patient request.

• B. A shorter antibiotic course is still unnecessary for a viral infection and does not resolve the underlying issue of inappropriate prescribing.
• C. Imaging and blood tests are not indicated for a straightforward viral upper respiratory infection and would not change management.
• D. Prescribing antibiotics for a viral illness provides no clinical benefit and promotes resistance; patient request alone does not justify it.
• E. Dismissing the request without explanation fails to address the patient's concerns or educate her about her condition.

A valid advance directive created while the patient had decision-making capacity reflects the patient's own autonomous wishes and takes legal and ethical precedence over the preferences of family members, even next of kin, once the patient loses capacity. The physician's duty is to honor the previously expressed wishes of the patient rather than defer to surrogates who wish to override them. Ethics consultation can be helpful for support or conflict resolution, but it is not required before honoring an unambiguous, validly executed directive, and a court order is not necessary.

• A. An advance directive specifically limits surrogate authority when it clearly addresses the situation at hand; it is not superseded by family preference.
• B. Family wishes cannot override a clear, validly executed advance directive expressing the patient's own wishes.
• C. No court order is needed to honor a legally valid advance directive.
• E. While ethics consultation can assist with family conflict, it is not a required prerequisite to honoring a clear advance directive.

Withdrawing life-sustaining treatment, such as mechanical ventilation, at the request of a patient or authorized surrogate is a well-established and legally protected practice, ethically distinct from physician-assisted suicide or euthanasia. The intent is to discontinue treatment that is no longer wanted and allow the underlying disease to take its natural course, not to directly cause death. Administering morphine to relieve dyspnea during this process is palliative symptom management under the principle of double effect, not an act of euthanasia, since the intent is comfort, not death. Physician-assisted suicide requires the physician to provide a lethal means for the patient to self-administer, which did not occur here.

• A. Assisted suicide requires the patient to self-administer a lethal agent; here the physician withdrew unwanted treatment and provided symptom relief.
• B. The surrogate validly consented on the patient's behalf based on her previously expressed wishes, so this is not involuntary.
• C. The intent of the morphine infusion was symptom control for dyspnea, not to actively end life, which distinguishes this from euthanasia.
• D. Withdrawal of life-sustaining treatment at the request of a patient or valid surrogate is legal and does not require judicial review.

The principle of double effect permits an action with a foreseeable harmful effect if the intended effect is good (pain relief), the harmful effect is not the means to the good effect, the action itself is not inherently wrong, and the good effect is proportionate to the risk of harm. Appropriately titrated opioids for severe pain in a terminally ill patient meet these criteria even if respiratory depression and hastened death are foreseeable. The physician's intent is analgesia, not euthanasia, and the dose is adjusted based on symptom need rather than an intent to cause death. This distinguishes ethically permissible palliative sedation/analgesia from physician-assisted death.

• A. Decision-making authority is not simply transferred by a family member's acceptance of risk when the patient himself is competent and present.
• C. Core ethical principles such as beneficence, nonmaleficence, and intent still apply in terminal illness; they are not suspended.
• D. General consent to treatment does not by itself justify an action; the ethical justification rests on intent and proportionality, not blanket consent.
• E. Hospice status does not create a special legal exemption; the ethical analysis is based on intent and proportionality, not setting.