RxReady

PEBC Biostatistics and Pharmacoeconomics Practice Questions

NNT, study design, bias, power and cost analyses: ten evidence-based practice questions for the PEBC MCQ.

Updated November 13, 2026. Written by RxReady to the 2026 PEBC blueprint. These are practice questions, not real or recalled exam questions.

Biostatistics and pharmacoeconomics questions reward a handful of formulas and definitions applied carefully. If you can calculate ARR, RRR and NNT, name the study design and spot the bias, you can win most of these marks.

This set covers crossover and factorial designs, statistical power, detection and recall bias, ARR, RRR, NNT and NNH, sensitivity, and cost-benefit analysis and study perspective.

Write the 2x2 table or the formula before you calculate. Most wrong answers come from mixing up absolute and relative risk.

Practice by topic

Ten biostatistics questions

Question 1 · Study design

For which of the following research questions would a crossover design be most suitable?

  1. AComparing two antibiotics for cure of community-acquired pneumonia
  2. BComparing two short-acting analgesics for chronic, stable osteoarthritis pain
  3. CComparing two vaccines for prevention of influenza
  4. DComparing two antiplatelet regimens for mortality after myocardial infarction
Show the answer

Answer: B) Comparing two short-acting analgesics for chronic, stable osteoarthritis pain

Crossover designs, in which each patient receives both treatments in sequence, require a chronic, stable condition and a treatment whose effect is reversible and short-lived (allowing a washout). Pneumonia cure is a one-time outcome that changes the disease state, vaccine immunity is long-lasting, and mortality is an irreversible outcome, so none can be studied by crossover.

Why the other options are wrong

  • A) Pneumonia cure is a one-time outcome that changes the disease, so patients cannot cross over.
  • C) Vaccine immunity lasts a long time, so the first vaccine would carry over into the second period.
  • D) Death is irreversible, so a patient cannot receive the second treatment period.

Key point: Crossover needs a chronic, stable disease and a short-acting, reversible treatment with a washout.

Question 2 · Pharmacoeconomics

Which of the following pharmacoeconomic analyses measures both costs and outcomes in monetary units?

  1. ACost-effectiveness analysis
  2. BCost-utility analysis
  3. CCost-minimization analysis
  4. DCost-benefit analysis
Show the answer

Answer: D) Cost-benefit analysis

Cost-benefit analysis converts health outcomes to dollars, allowing results to be expressed as a net benefit or benefit-to-cost ratio, and allowing comparison of programs with different outcomes. Cost-effectiveness analysis uses natural units (e.g., mmHg, life-years), cost-utility analysis uses QALYs, and cost-minimization assumes equivalent outcomes and compares only costs.

Why the other options are wrong

  • A) Cost-effectiveness analysis uses natural units for outcomes.
  • B) Cost-utility analysis uses QALYs.
  • C) Cost-minimization assumes equal outcomes and compares costs only.

Key point: Cost-benefit = everything in dollars.

Question 3 · Biostatistics

A trial is designed with alpha = 0.05 and beta = 0.20. Which of the following is the statistical power of the trial?

  1. A5%
  2. B20%
  3. C80%
  4. D95%
Show the answer

Answer: C) 80%

Power = 1 - beta = 1 - 0.20 = 0.80 (80%), the probability of detecting a true difference of the specified size. Alpha (5%) is the acceptable probability of a type I error, and beta (20%) is the probability of a type II error.

Why the other options are wrong

  • A) 5% is alpha, the accepted type I error risk, not power.
  • B) 20% is beta, the type II error risk, not power.
  • D) 95% is 1 - alpha, the confidence level, not power.

Key point: Power = 1 - beta.

Question 4 · Biostatistics

Which of the following changes in study planning would increase the statistical power of a trial, all else being equal?

  1. AChoosing an outcome measure with greater variability
  2. BExpecting a smaller difference between treatments
  3. CLowering alpha from 0.05 to 0.01
  4. DIncreasing the sample size
Show the answer

Answer: D) Increasing the sample size

Power (1 - beta) increases with a larger sample size, a larger expected effect size, a higher alpha and less variability. A more variable outcome measure, a stricter alpha (0.01) and targeting a smaller effect all reduce power for a given design.

Why the other options are wrong

  • A) More variability in the outcome lowers power, making a real effect harder to detect.
  • B) Looking for a smaller effect lowers power for the same sample size.
  • C) A stricter alpha (0.01) lowers power, raising the chance of a type II error.

Key point: Power rises with larger sample size, larger effect size, higher alpha and less variability.

Question 5 · Pharmacoeconomics

Which of the following costs is counted in a pharmacoeconomic analysis only when it takes the societal perspective rather than the public payer perspective?

  1. ADrug acquisition and dispensing costs
  2. BLost productivity from work absence
  3. CPhysician fees for office visits
  4. DHospitalization and emergency visit costs
Show the answer

Answer: B) Lost productivity from work absence

The societal perspective is the broadest and includes indirect costs, such as productivity losses from illness or caregiving, in addition to direct medical costs. A public payer (ministry of health) perspective includes only the direct costs borne by the payer, such as drugs, physician services, hospital care and laboratory tests.

Why the other options are wrong

  • A) Drug and dispensing costs are direct costs counted by the public payer.
  • C) Physician fees are direct costs paid by the public payer.
  • D) Hospital and emergency costs are direct costs paid by the public payer.

Key point: Societal perspective = payer costs + indirect costs (lost productivity, caregiving).

Question 6 · Study design

In one RCT, participants were randomized to ASA or placebo and, separately, to vitamin E or placebo, so that each participant fell into one of four groups. Which of the following designs was used?

  1. A2 x 2 factorial design
  2. BCrossover design
  3. CCluster randomized design
  4. DNon-inferiority design
Show the answer

Answer: A) 2 x 2 factorial design

A factorial design randomizes participants to two or more interventions simultaneously, allowing each to be evaluated (and their interaction examined) within the same trial; a 2 x 2 factorial produces four groups. In a crossover design each participant receives treatments sequentially, and cluster designs randomize groups rather than individuals.

Why the other options are wrong

  • B) In a crossover, each patient gets treatments one after another, not two randomizations at once.
  • C) Cluster designs randomize groups (e.g., clinics), not individual participants.
  • D) Non-inferiority describes the hypothesis tested, not a four-group design.

Key point: 2 x 2 factorial: two separate randomizations in one trial, giving four groups.

Question 7 · Bias

An open-label RCT compared two treatments for chronic insomnia, with the primary outcome being patient-rated sleep quality. Which of the following is the greatest threat to the internal validity of this trial?

  1. AHealthy-user bias
  2. BRecall bias from interviewing cases and controls about past sleep exposures
  3. CLead-time bias
  4. DBias from knowing the assigned treatment when rating a subjective outcome
Show the answer

Answer: D) Bias from knowing the assigned treatment when rating a subjective outcome

When participants and assessors know the assigned treatment, their expectations can influence subjective outcomes such as sleep quality (performance and detection bias); blinding is most important when outcomes are subjective. Recall bias applies to case-control designs, lead-time bias to screening, and healthy-user bias to observational studies.

Why the other options are wrong

  • A) Healthy-user bias affects observational studies of preventive therapy, not an RCT of insomnia.
  • B) Recall bias applies to case-control interviews, not to this RCT.
  • C) Lead-time bias relates to screening and earlier diagnosis, not insomnia treatment.

Key point: Blinding matters most when the outcome is subjective.

Question 8 · ARR, RRR, NNT, NNH

In a 3-year randomized trial, the primary outcome occurred in 10% of patients receiving placebo and 7.5% of patients receiving the study drug. Which of the following is the number needed to treat (NNT) over 3 years?

  1. A4
  2. B40
  3. C10
  4. D13
Show the answer

Answer: B) 40

ARR = 10% − 7.5% = 2.5% (0.025). NNT = 1 ÷ 0.025 = 40. Why not the others: A: This uses the relative risk reduction (25%) instead of the absolute difference. C: This uses 1 ÷ 0.10, the placebo event rate. D: This uses 1 ÷ 0.075, the event rate in the drug group. Key takeaway: NNT = 1 ÷ ARR (use the absolute, not relative, difference) for the trial's time frame.

Why the other options are wrong

  • A) 4 uses the relative risk reduction (25%) instead of the absolute difference.
  • C) 10 is 1 / 0.10, the placebo event rate, not the risk difference.
  • D) 13 is 1 / 0.075, the drug group event rate, not the risk difference.

Key point: NNT = 1 / ARR (absolute, not relative, difference) for the trial's time frame.

Question 9 · Bias

Investigators interviewed mothers of infants born with a congenital malformation and mothers of healthy infants about medications taken during pregnancy. Which of the following biases is of greatest concern in this study?

  1. ARecall bias
  2. BAttrition bias
  3. CLead-time bias
  4. DHealthy-user bias
Show the answer

Answer: A) Recall bias

In case-control studies that rely on interviews, mothers of affected infants tend to search their memories more thoroughly for possible exposures than mothers of healthy infants, producing differential recall (recall bias) that can exaggerate an association. Lead-time bias relates to screening, attrition bias to loss to follow-up in prospective studies, and healthy-user bias to preventive-therapy cohorts.

Why the other options are wrong

  • B) Attrition bias comes from loss to follow-up in prospective studies; this is a one-time interview.
  • C) Lead-time bias relates to screening and earlier diagnosis, not to recalling past exposures.
  • D) Healthy-user bias affects cohorts of people using preventive therapy, not interview-based case-control studies.

Key point: Interview-based case-control studies, especially of birth defects, are prone to recall bias.

Question 10 · Biostatistics

A clinician wants a test that, when negative, can reliably rule out a serious condition. Which characteristic should the test have?

  1. ALow negative predictive value
  2. BHigh specificity
  3. CHigh positive predictive value
  4. DHigh sensitivity
Show the answer

Answer: D) High sensitivity

A highly sensitive test has few false negatives, so a negative result effectively rules the condition out ("SnNout"). A highly specific test has few false positives, so a positive result helps rule the condition in ("SpPin").

Why the other options are wrong

  • A) A low NPV means negative results are often wrong, the opposite of what is needed.
  • B) High specificity helps rule a condition in with a positive result (SpPin), not out.
  • C) High PPV makes positive results reliable; it does not help rule out with a negative test.

Key point: SnNout: a negative highly Sensitive test rules OUT; SpPin: a positive Specific test rules IN.

Find your weakest topics

The free diagnostic grades you across all six PEBC competencies, then builds a study plan around the areas costing you the most marks. We also have 2,000+ practice questions and 98 OSCE stations where you talk to an AI patient out loud.

Take the free diagnosticTry a free OSCE station

More free study guides