A trial analyses time until a first heart attack in two groups, with some patients still event-free at the end of the study. Which method is most suitable?

Kaplan-Meier survival analysis with a log-rank test

A chi-square test of the final proportion of events only

Which study design provides the strongest evidence for the effect of a treatment, in a single study?

A cross-sectional population survey

Biostatistics and evidence-based practice

33 Gulf Prometric Pharmacist (DHA, DOH, MOH, SCFHS, QCHP) practice questions on Biostatistics and evidence-based practice, part of Pharmacy Practice, Law and Management. The first 25 are below; the first 5 with answers and explanations (tap Show answer after you have chosen), the rest in Learn with a free account.

Question 1 · Medium
The authors of a drug trial conclude that the drug works, when in truth it has no effect. Which error have they made?
1. AConfounding bias
2. BType I error
3. CLead-time bias
4. DType II error
Show answerHide answer
Answer: B · Type I error
Concluding that an effect exists when the null hypothesis is true is a false positive, or type I (alpha) error. Alpha is usually set at 0.05, the accepted probability of this error.
- A: Confounding is a systematic distortion by a third factor, not a decision error.
- C: Lead-time bias relates to screening and apparent survival time.
- D: A type II error is missing a real effect (a false negative).
Remember this
Type I error is a false positive; type II is a false negative.
Question 2 · Medium
In a study of coffee and heart disease, smokers drink more coffee and also have more heart disease. Smoking here is a:
1. ARandom error
2. BPlacebo
3. COutcome
4. DConfounder
Show answerHide answer
Answer: D · Confounder
A confounder is related to both the exposure and the outcome and is not on the causal pathway. It can create or hide an association. It is handled by randomisation, restriction, matching or adjustment.
- A: Random error is chance variation, not a systematic factor.
- B: A placebo is an inert comparator.
- C: The outcome is heart disease.
Remember this
A confounder is linked to both exposure and outcome and is not on the causal path.
Question 3 · Medium
A trial analyses time until a first heart attack in two groups, with some patients still event-free at the end of the study. Which method is most suitable?
1. AKaplan-Meier survival analysis with a log-rank test
2. BA simple range
3. CA chi-square test of the final proportion of events only
4. DA paired t-test
Show answerHide answer
Answer: A · Kaplan-Meier survival analysis with a log-rank test
Time-to-event data with censored participants are analysed using survival methods. Kaplan-Meier curves describe survival, the log-rank test compares them, and Cox regression gives a hazard ratio.
- B: The range does not compare groups.
- C: A final proportion ignores the timing of events.
- D: A paired t-test needs paired, uncensored measurements.
Remember this
Time-to-event data with censoring: Kaplan-Meier curves, log-rank test, Cox hazard ratio.
Question 4 · Easy
A new assay detects 95 of the 100 patients known to have a condition. Which test property does the figure of 95% describe?
1. ASpecificity
2. BNegative predictive value
3. CPositive predictive value
4. DSensitivity
Show answerHide answer
Answer: D · Sensitivity
Sensitivity = true positives / (true positives + false negatives) = 95 / 100 = 95%. A highly sensitive test rarely misses disease, so a negative result helps to rule it out. Specificity and the predictive values use other denominators.
- A: Specificity relates to the healthy people who test negative.
- B: NPV is the share of negative results that are truly negative.
- C: PPV is the share of positive results that are true positives.
Remember this
Sensitivity = true positives / all who have the disease.
Question 5 · Easy
Which study design provides the strongest evidence for the effect of a treatment, in a single study?
1. AA cross-sectional population survey
2. BA randomised controlled trial
3. CA case report
4. DA case series
Show answerHide answer
Answer: B · A randomised controlled trial
Random allocation balances known and unknown confounders between groups, so differences in outcome can be attributed to the treatment. Blinding further reduces bias.
- A: A cross-sectional survey measures exposure and outcome at one time and cannot show effect.
- C: A case report has no comparison group.
- D: A case series has no control group.
Remember this
A well-run randomised controlled trial gives the strongest single-study evidence for treatment effect.
Question 6 · Easy
In a cohort study, participants are grouped according to their:
1. ARandom numbers assigned by the investigator
2. BExposure status, then followed for the outcome
3. CAge only
4. DOutcome status, then asked about past exposure
Answer and explanation in LearnFree account
Question 7 · Hard
In a trial, serious bleeding occurs in 3% of patients on a new anticoagulant and 1% on the comparator. What is the number needed to harm, and what does it mean?
1. A2; one extra bleed for every 2 patients treated
2. B50; one extra bleed for every 50 patients treated
3. C33; one extra bleed for every 33 patients treated
4. D100; one extra bleed for every 100 patients treated
Answer and explanation in LearnFree account
Question 8 · Medium
In a normal distribution, about what proportion of values lie within 2 standard deviations of the mean?
1. AAbout 50%
2. BAbout 95%
3. CAbout 68%
4. DAbout 99.7%
Answer and explanation in LearnFree account
Question 9 · Hard
In a trial, the event rate is 20% in the control group and 15% in the treatment group. What is the NNT?
1. A5
2. B20
3. C4
4. D25
Answer and explanation in LearnFree account
Question 10 · Medium
The statistical power of a study is:
1. AThe size of the effect between the two groups
2. BThe probability of detecting a true effect, equal to 1 minus type II error
3. CThe same as the p-value
4. DThe probability of a false positive finding when there is no effect
Answer and explanation in LearnFree account
Question 11 · Hard
The risk of an outcome is 8% in the control group and 6% in the treated group. What are the relative risk reduction (RRR) and the absolute risk reduction (ARR)?
1. ARRR 2%, ARR 25%
2. BRRR 75%, ARR 2%
3. CRRR 25%, ARR 2%
4. DRRR 25%, ARR 25%
Answer and explanation in LearnFree account
Question 12 · Medium
A meta-analysis reports an I-squared of 85%. What does this indicate?
1. APublication bias has been excluded
2. BThe studies are all identical
3. CSubstantial heterogeneity between the included studies
4. DThe pooled effect is statistically significant and clinically important
Answer and explanation in LearnFree account
Question 13 · Medium
In a case-control study, mothers of children with a birth defect remember their drug use in pregnancy better than mothers of healthy children. This is an example of:
1. AAttrition bias
2. BRecall bias
3. CPublication bias
4. DLead-time bias
Answer and explanation in LearnFree account
Question 14 · Easy
An assay correctly gives a negative result in 190 of 200 healthy volunteers. Which test property does 95% describe?
1. APositive predictive value
2. BNegative predictive value
3. CSpecificity
4. DSensitivity
Answer and explanation in LearnFree account
Question 15 · Medium
A study reports a correlation coefficient of -0.8 between daily sodium intake and bone density. Which interpretation is correct?
1. AA strong inverse association exists, but it does not prove causation
2. BBone density rises as sodium rises
3. CSodium causes a decrease in bone density
4. DThere is no association between sodium intake and bone density
Answer and explanation in LearnFree account
Question 16 · Medium
A funnel plot of the studies in a meta-analysis shows that small studies with negative results are missing, and the plot is asymmetrical. This suggests:
1. APossible publication bias
2. BThat randomisation was not used
3. CPerfect homogeneity between studies
4. DA very large sample in each study
Answer and explanation in LearnFree account
Question 17 · Hard
A trial of 100 participants gives a 95% confidence interval of a certain width. Other things equal, how many participants are needed to halve the width of the interval?
1. A400
2. B800
3. C300
4. D200
Answer and explanation in LearnFree account
Question 18 · Medium
A study reports a relative risk of 0.85 with a 95% confidence interval of 0.70 to 1.04. What is the best conclusion?
1. AThe study has no uncertainty
2. BThe treatment is significantly harmful
3. CThe result is not statistically significant, as the interval includes 1
4. DThe treatment is significantly effective, since the estimate is below 1
Answer and explanation in LearnFree account
Question 19 · Easy
The number needed to treat (NNT) is calculated as:
1. AThe odds ratio divided by the risk ratio
2. B1 divided by the relative risk
3. CThe relative risk reduction multiplied by 100
4. D1 divided by the absolute risk reduction
Answer and explanation in LearnFree account
Question 20 · Medium
Which test is appropriate for comparing the proportion of patients who respond between two treatment groups?
1. AUnpaired t-test of the two group means
2. BOne-way ANOVA across the group means
3. CPearson correlation coefficient test
4. DChi-square test (or Fisher's exact test)
Answer and explanation in LearnFree account
Question 21 · Medium
A trial reports a relative risk of 0.60 for stroke with a new drug compared with placebo. How should this be interpreted?
1. AThe drug had no effect
2. BThe drug reduced the risk of stroke by 40% relative to placebo
3. CThe drug increased the risk by 60%
4. DThe drug reduced the absolute risk of stroke by 60 percentage points
Answer and explanation in LearnFree account
Question 22 · Medium
A researcher fixes the numbers of cases and controls before collecting exposure data. Which measure of association fits, and why?
1. AHazard ratio, because the time to the event was recorded
2. BOdds ratio, because incidence in exposed and unexposed people cannot be calculated
3. CRelative risk, because the groups were followed forward over time
4. DAttributable risk, because exposure was allocated at random
Answer and explanation in LearnFree account
Question 23 · Hard
In a case-control study, 40 of 100 cases and 20 of 100 controls were exposed to a drug. What is the odds ratio?
1. A0.37
2. B1.50
3. C2.67
4. D2.0
Answer and explanation in LearnFree account
Question 24 · Medium
In a double-blind trial, neither the participants nor the investigators know the treatment allocation. The main purpose is to reduce:
1. AConfounding by age and sex distribution
2. BRandom error
3. CPerformance and detection bias
4. DAttrition
Answer and explanation in LearnFree account
Question 25 · Easy
A p-value of 0.03 in a trial means that:
1. AThere is a 3% chance that the null hypothesis is true
2. BThe effect is clinically important
3. CThere is a 97% probability that the drug truly works in this population
4. DIf there were no true effect, a result this extreme would occur about 3% of the time
Answer and explanation in LearnFree account

8 more Biostatistics and evidence-based practice questions

Answer them one by one, keep your score and see your weak topics in Learn. Signed-out visitors try 25 per exam; sign in with Google to see them all.

Study Pharmacy Practice, Law and Management

Biostatistics and evidence-based practice

The authors of a drug trial conclude that the drug works, when in truth it has no effect. Which error have they made?

In a study of coffee and heart disease, smokers drink more coffee and also have more heart disease. Smoking here is a:

A trial analyses time until a first heart attack in two groups, with some patients still event-free at the end of the study. Which method is most suitable?

A new assay detects 95 of the 100 patients known to have a condition. Which test property does the figure of 95% describe?

Which study design provides the strongest evidence for the effect of a treatment, in a single study?

In a cohort study, participants are grouped according to their:

In a trial, serious bleeding occurs in 3% of patients on a new anticoagulant and 1% on the comparator. What is the number needed to harm, and what does it mean?

In a normal distribution, about what proportion of values lie within 2 standard deviations of the mean?

In a trial, the event rate is 20% in the control group and 15% in the treatment group. What is the NNT?

The statistical power of a study is:

The risk of an outcome is 8% in the control group and 6% in the treated group. What are the relative risk reduction (RRR) and the absolute risk reduction (ARR)?

A meta-analysis reports an I-squared of 85%. What does this indicate?

In a case-control study, mothers of children with a birth defect remember their drug use in pregnancy better than mothers of healthy children. This is an example of:

An assay correctly gives a negative result in 190 of 200 healthy volunteers. Which test property does 95% describe?

A study reports a correlation coefficient of -0.8 between daily sodium intake and bone density. Which interpretation is correct?

A funnel plot of the studies in a meta-analysis shows that small studies with negative results are missing, and the plot is asymmetrical. This suggests:

A trial of 100 participants gives a 95% confidence interval of a certain width. Other things equal, how many participants are needed to halve the width of the interval?

A study reports a relative risk of 0.85 with a 95% confidence interval of 0.70 to 1.04. What is the best conclusion?

The number needed to treat (NNT) is calculated as:

Which test is appropriate for comparing the proportion of patients who respond between two treatment groups?

A trial reports a relative risk of 0.60 for stroke with a new drug compared with placebo. How should this be interpreted?

A researcher fixes the numbers of cases and controls before collecting exposure data. Which measure of association fits, and why?

In a case-control study, 40 of 100 cases and 20 of 100 controls were exposed to a drug. What is the odds ratio?

In a double-blind trial, neither the participants nor the investigators know the treatment allocation. The main purpose is to reduce:

A p-value of 0.03 in a trial means that: