1. When producing evidence-based guidelines, grading the level of evidence is now regarded as standard practice by guideline producing organisations. What is the key principle underlying this practice? Select one: * It allows funding bodies to prioritise and commission the most effective treatments * It allows users of the guideline to evaluate the best treatments * It enables valid statistical comparisons between different RCTs of similar interventions * It helps to define the most cost-effective treatments * It is derived from the impact factor of the medical journal which published the original study 2. In Uganda, a researcher wants to assess the effect of treated mosquito nets on reducing the risk of malaria in a community. They have access to a large district in northern Uganda which includes several remote villages. Preliminary data indicates that individual nets in a community have limited benefit and it requires the majority of households to use them to see a potentially appreciable difference. What is the best way to see if treated nets reduce the risk ut malaria? Select one: * Case control study * Cluster randomised control trial * Cohort study * Crossover randomised control trial
Factorial randomised control trial 3. A high-quality randomised-controlled trial investigating the effect of a new treatment of ovarian cancer shows that the remission rate at one year is similar to usual care, the P-value = 0.4. This means which of the following? Select one: * Both treatments are effective * It is not possible to determine whether one treatment is better than the other using this information alone * Neither treatment is effective * The best estimate of the treatment effect is 0.4 * The statistical power of this study is 60% 4. There has been much debate about the potential risk of insulin therapy in the causation of cancer. This debate has focused on results of observational studies and randomised-controlled trials (RCTs). Considering the grading of evidence, in which specific setting can observational studies sometimes provide stronger evidence than RCTs? Select one: * In the evaluation of rare adverse effects of a given therapy * When injection-based therapies are under review, given ethical concerns about participants in an RCT receiving placebo injections over an extended period * When observational studies contain more than 10-20 fold the number of participants in an RCT * When the end-point in question is very clearly defined eg. mortality * When the observational study is conducted over an extended period of time 5. What term is defined by the following statement: "The variability of the data set considering the deviation of each data point and distance from mean"? Select one: * ANOVA * Chi-squared * Standard deviation * Standard error of the mean • Mathematical mean 6. A specialist tertiary rheumatology clinic has recently introduced a new website detailing its clinic schedule and staff list and also providing patient information leaflets and links. Under which aspect, if any, of clinical governance is this most appropriately defined for the clinic? Select one: * Clinical äudit * Clinical effectiveness programmes * Clinical risk management * Not a component of clinical governance * Patieft and public involvement 7. A practice produced and presented an audit of its Medicines Use Reviews (MURs) in conjunction with a pharmacy for patients with hyperthyroidism. From the audit a few points were derived, one of which was regarding patients' awareness of the adverse events associated with these drugs. Which of the following would be the next stage in the audit process? Select one: * Data analysis * Data collection * Identifying standards • Implementing change Needs assessment 8. A study has a small sample size (N = 5 in two distinct cohorts). What must we be confident with to overcome this small sample size? Select one: * That the data are normally distributed and require non-parametric statistics * That the data are not normally distributed and require parametric statistics * That the power of the test demonstrates that 10 subjects elicits >80% power to determine differences (P<0.05) only amongst groups * That the power of the test demonstrates that 10 subjects elicits >90% power to determine differences (P<0.05) within the methodology * That the power of the test demonstrates that 10 subjects elicits <95% power 9. If a longitudinal study estimated the relative risk of colorectal cancer over 25 years in relation to dietary consumption in a sample of participants who were all apparently free of colorectal cancer at baseline what type of study would it be? Select one: Case-control study Cluster randomised-controlled trial Ecological study Prospective cohort study
Retrospective cohort study

Here are the answers with explanations:

1. Key principle underlying grading of evidence in guidelines
It allows users of the guideline to evaluate the best treatments
Grading evidence (e.g., GRADE system) communicates the confidence/certainty behind each recommendation, so clinicians and other guideline users can judge how much weight to place on each recommendation and make informed decisions.

2. Best study design for mosquito nets and malaria in Uganda
Cluster randomised control trial
The preliminary data explicitly states that individual nets have limited benefit - the intervention works at the community (cluster) level. Randomising entire villages (clusters) to receive treated nets or not is the appropriate design. This accounts for the "herd protection" effect and avoids contamination between individuals within the same village.

3. P-value = 0.4 in the ovarian cancer RCT
It is not possible to determine whether one treatment is better than the other using this information alone
A p-value of 0.4 means the observed difference could easily occur by chance (it is not statistically significant), but it does NOT prove equivalence. Without knowing the confidence interval, sample size, and power, you cannot conclude the treatments are equal. The p-value is not the treatment effect, not the power, and says nothing about absolute efficacy.

4. When can observational studies provide stronger evidence than RCTs?
In the evaluation of rare adverse effects of a given therapy
RCTs are typically powered to detect efficacy outcomes and are too small and too short to detect rare adverse events. Large observational studies (pharmacovigilance, registry data, case-control studies) can include far more patients over much longer follow-up, making them better suited to detecting rare harms - this is exactly why the insulin-cancer debate was driven by observational data.

5. Term defined as "variability of the dataset considering deviation of each data point from the mean"
Standard deviation
Standard deviation (SD) is the square root of the variance, which is the average of squared deviations from the mean. It quantifies spread/variability of individual data points around the mean.

6. Clinical governance domain for the rheumatology clinic website
Patient and public involvement
The website provides patient information leaflets, clinic schedules, and links - its primary function is informing and engaging patients and the public. This falls under patient and public involvement (PPI), one of the seven pillars of clinical governance.

7. Next stage in the audit cycle after deriving findings
Implementing change
The audit has already gone through: needs assessment → identifying standards → data collection → data analysis → presenting results. The finding about patients' lack of awareness of adverse events is an audit result, so the next step in the audit cycle is to implement change (e.g., improve patient counselling or information provision). The cycle then re-audits to close the loop.

8. What must we be confident with to overcome a small sample size (N=5)?
That the power of the test demonstrates that 10 subjects elicits >80% power to determine differences (P<0.05) only amongst groups
With a very small sample size, the key concern is statistical power - the ability to detect a true difference if one exists. The conventional threshold is 80% power at P<0.05. You need to demonstrate that even with N=10 total (5 per group), the power calculation supports the study's ability to detect meaningful differences.

9. Study type: longitudinal, 25 years, dietary consumption, colorectal cancer-free at baseline
Prospective cohort study
Participants are enrolled free of disease at baseline, then followed forward in time (25 years) to observe who develops colorectal cancer. Exposure (diet) is assessed before the outcome occurs. This is the defining feature of a prospective cohort study. It is not retrospective because follow-up goes forward from baseline, and it is not a case-control because cases are not selected after the fact.

11. A recent meta-analysis reported that drug X compared to standard therapy was associated with increased cardiovascular mortality. Within the meta-analysis the number of events/total treated for drug X and controls were 86/14376 and 72/11644 respectively. What is the odds ratio for cardiovascular events with drug X compared to control? Select one: 0.0005 0.60 0.86 0.97 1.20 12. A new urine test (NIC-UR) has been developed to identify the presence of nicotine in the urine of suspected smokers. This was tested in a sample of 300 volunteers. NIC-UR tested positive in 120 of 150 smokers and positive in 5 of 150 non-smokers. What is the sensitivity of this test? Select one: * 20%
40%
80%
97% * 100% 13. A team of health economists perform a study that examines blood flow responses to a fall in temperature and then the addition of nitrate therapy in Raynaud's patients. They collect blood flow responses in 30 female participants at 3 time-points: (1) baseline; (2) when cold and (3) with nitrate use, The data are normally distributed. What would be the most appropriate statistical approach to test the null-hypothesis that blood flow at each time point is not different? Select one: Multiple logistic regression Multiple paired sample t-tests Pearson product moment correlation coefficient Repeated measures ANOVA
Spearman rank-order correlation coefficient 14. Which of the following is an advantage of a case-control study? Select one: * Bias in the assessment of exposure to the factor of interest is minimised * Dependence on recall by subjects in the study is minimised * It is possible to determine the true incidence of the disease * It may be used to study the etiology of a rare disease * Multiple disease outcomes following a selected exposure can be readily studied 15. If a recommendation on a new treatment strategy was published in a guidance document which depends upon evidence published by a research group whose principal investigator was also on the advisory board of the company manufacturing the product, how would this be graded? Select one: • Grade A • Grade B • Grade C • Grade D • No grading provided 16. A PhD student is investigating the impact of exercise on dementia risk in the elderly (60 years old and above). He examines the database of a local primary care practice to identify patients with dementia 60-years and over. Within the database he identified 35 patients with dementia from a total of 2073 patients who were 60 years or older. What is the prevalence of dementia within the practice sample? Select one: * 0.86% * 0.97%
1.7%
2.6%
3.2% 18. Which of the following is TRUE about incidence and prevalence? Select one: If the incidence is decreasing the prevalence is also decreasing • Incidence and prevalence are Interchangeable terms Incidence is reported in raw numbers while prevalence is reported as a ratio Prevalence is used to describe the rate at which new cases occur • The incidence for a given disease can be low at the same time that its prevalence can be high 19. A qualitative study is being designed to investigate patients' perceptions of using lip fillers. What will the methodology involve? Select one: Analysis of data using statistical methodology e.g. chi-squared testing • An inductive approach with purposeful sampling, generating emergent themes Double blind folding of participants Focus groups and analysis of data with coding followed by a T-test •Randomisation of participants into 2 groups 20. A junior doctor is interested in the association between psoriasis and metabolic syndrome. They undertake a study to investigate the association between psoriasis area and severity index (PASI) score and the following cardiovascular risk factors: cholesterol, HDL, fasting glucose, systolic BP and diastolic BP. They perform a simple bivariate regression analysis in a non-random sample of 20 volunteers and obtain the following correlation coefficients: Systolic BP (r = 0.54, p = 0.62) Diastolic BP (r = 0.420, p = 0.72) Cholesterol (г = 0.08, р = 0.81) HdL (r = -0.85, p = 0.01) fasting glucose (r = -0.56, p = 0.03) Which one of the risk factors was most closely associated with PASI? Select one: Cholesterol Diastolic BP Fasting glucose • HDL Systolic BP

Here are the answers with full explanations:

11. Odds Ratio for cardiovascular events with drug X vs control
Answer: 0.97
Calculate as follows:
  • Drug X: events = 86, no events = 14376 - 86 = 14290
  • Control: events = 72, no events = 11644 - 72 = 11572
Odds of event with drug X = 86/14290 = 0.006018 Odds of event with control = 72/11572 = 0.006221
OR = 0.006018 / 0.006221 = 0.968 ≈ 0.97
The OR is just below 1.0, meaning drug X has very slightly lower odds of cardiovascular events than control - but this difference is negligible and not meaningful.

12. Sensitivity of NIC-UR test
Answer: 80%
Sensitivity = True Positives / (True Positives + False Negatives) = 120 / 150 (smokers who tested positive / all smokers) = 80%
(Specificity = 145/150 = 97% - the 5 false positives among 150 non-smokers)

13. Most appropriate statistical test for blood flow at 3 time-points, normally distributed, same participants
Answer: Repeated measures ANOVA
Three key features drive this choice:
  1. Same participants measured at multiple time-points (repeated/related measures)
  2. Three time-points (more than 2, so a t-test is insufficient)
  3. Normally distributed data (parametric test appropriate)
Repeated measures ANOVA compares means across 3+ related time-points in the same subjects. Multiple paired t-tests would inflate the Type I error rate (multiple comparisons problem).

14. Advantage of a case-control study
Answer: It may be used to study the etiology of a rare disease
Case-control studies start by selecting people with the disease (cases) and work backwards to examine exposure. This makes them efficient for rare diseases - you don't need to follow thousands of people waiting for a rare outcome to occur. The other options are disadvantages, not advantages (recall bias is a major weakness of case-control studies, and you cannot calculate true incidence from them).

15. Grading of a recommendation with conflicts of interest
Answer: Grade D
In most grading frameworks (e.g., SIGN, GRADE), Grade D represents the lowest level of evidence - typically expert opinion, or evidence from studies with significant methodological concerns. A conflict of interest (principal investigator on the advisory board of the manufacturer) severely undermines the validity and independence of the evidence, warranting the lowest grade. Some frameworks might say "no grading," but Grade D is the most appropriate formal answer here.

16. Prevalence of dementia in the practice sample
Answer: 1.7%
Prevalence = (Number with condition / Total population) × 100 = (35 / 2073) × 100 = 1.69% ≈ 1.7%

18. TRUE statement about incidence and prevalence
Answer: The incidence for a given disease can be low at the same time that its prevalence can be high
This is true for chronic, long-lasting diseases (e.g., diabetes, HIV on treatment). Even if few new cases arise each year (low incidence), the pool of existing cases accumulates over time because patients live with the disease for years (high prevalence). The other options are all false - they are not interchangeable, prevalence is not a rate of new cases, and falling incidence does not immediately reduce prevalence.

19. Qualitative study methodology for lip filler perceptions
Answer: An inductive approach with purposeful sampling, generating emergent themes
Qualitative research is characterised by:
  • Inductive reasoning - themes emerge from the data rather than being pre-specified
  • Purposeful sampling - participants are selected because they have relevant experience, not randomly
  • Thematic analysis - coding data to identify patterns/themes
  • No statistical testing, no randomisation, no blinding

20. Risk factor most closely associated with PASI score
Answer: HDL
The question asks for the factor most closely associated - this is determined by the correlation coefficient (r), not the p-value. The strength of association is the absolute value of r:
Factor|r|p-value
Systolic BP0.540.62
Diastolic BP0.420.72
Cholesterol0.080.81
HDL0.850.01
Fasting glucose0.560.03
HDL has the highest absolute correlation (r = -0.85), meaning it has the strongest linear relationship with PASI score. It is also statistically significant (p = 0.01). The negative sign simply means higher PASI is associated with lower HDL (a clinically logical finding).

21. Meta-analysis is a statistical technique which involves the combination of results of several studies that address a set of related research hypotheses, Meta-analyses are often important components of a systematic review used to determine the level and strength of evidence. Which of the following plots is commonly used to assess the risk of publication bias in a meta-analysis? Select one: * Box and whisker plot * Forest plot * Funnel plot * P-P (probability-probability) plot
Scatter plot 22. A mass screening programme was undertaken in London for colorectal cancer using faecal occult blood tests. Two rounds were taken 3 years apart, with histological confirmed diagnosis following colonoscopy; Screening Round FOB Positive histology positive 94, histology negative 826 FOB Negative Histology Positive14, histology negative 10259 What is the specificity of the test? Histology Negative Select one: * 10.2 * 87.0
91.9
92.6
99.9
 23. A double blind randomised trial compared the efficacy of a new psoriasis biologic against placebo. The end point in the experiment is the relative lowering in psoriasis area and severity index (PASI) score after 6 months. The statistical analysis of the mean difference between groups produced a p-value of 0.01, What does a p-value of 0.01 signify in this scenario? Select one: * 0,1% probability that the difference in means was due to chance * 0.1% probability that the means are similar in the two groups * 1% probability of a type 2 statistical error * 1% probability that the difference in means was due to chance * 5% probability that the difference in means was due to chance 24. What does the standard error of the mean indicate? Select one: * If the value is small, that a type 2 error has occurred * The amount of error in the prediction of the population mean * The deviation of the mean about the median value * The value above which the mean yields significance * The variability of the data set considering the deviation of each data point and distance from the mean 25. When organisations produce clinical guidelines, the level of evidence is often graded by reviewers in terms of quality. During the review process of a randomised controlled trial, which of the following factors would be most likely to lead to a downgrading of the quality rating? Select one: • The primary outcome measure was a surrogate marker of outcome rather than a major clinical event
The study participants were confined to one specific ethnic group The study was terminated earlier than planned There was evidence of inadequate concealment of treatment allocation in a double-blinded trial There were multiple substantial amendments to the study protocol during the course of the trial 26. In a GP practice of 10,000 patients, there are 10 new cases of breast cancer per year, with a total of 500 cases in the records. What are the incidence and prevalence? Select one: * Incidence 0.001/year and prevalence 0.05 * Incidence 0.01/year and prevalence 0.5
Incidence 0.0001/year and prevalence 0.005
Incidence 0.002/year and prevalence 0.10
Incidence 0.003/year and prevalence 0.08 27. When designing a study to determine whether there is an association between a particular exposure and an outcome, it is important to anticipate potential alternative explanations. Which of the following is a way of trying to control confounding?Select one: Blinding Post-hoc blocking * Randomisation
Screening * Validation 29. A study is carried out in which children are randomly assigned to receive either a newly formulated vaccine or the currently most commonly used vaccine, Participants are followed to monitor bath for side effects and for the effectiveness of each vaccine. What type of study is this?Select one: * Case-control study * Clinical trial * Ecological study * Prospective cohort study * Time-series study 30. A GP is considering an audit on the use of amitriptyline in patients to assess whether the appropriate safety measures and screening processes have taken place prior to treatment. Which of the following is the next step in the audit process? Select one: * Data analysis * Data collection * Identify standards
Implement change
Needs assessment

Here are the answers with full explanations:

21. Plot used to assess publication bias in meta-analysis
Answer: Funnel plot
A funnel plot displays each study's effect size against its standard error (or sample size). In the absence of publication bias, studies scatter symmetrically around the pooled estimate in an inverted funnel shape. Asymmetry suggests smaller negative studies are missing (i.e., publication bias). A Forest plot displays individual study results and the pooled estimate, but does not assess publication bias.

22. Specificity of the FOB test
Answer: 92.6
First, set up the 2×2 table:
Histology +Histology -
FOB +94 (TP)826 (FP)
FOB -14 (FN)10,259 (TN)
Specificity = TN / (TN + FP) = 10,259 / (10,259 + 826) = 10,259 / 11,085 = 0.9255 = 92.6%
(Sensitivity = 94 / (94 + 14) = 94/108 = 87.0% - note 87.0 is the sensitivity, not specificity)

23. What does p = 0.01 signify?
Answer: 1% probability that the difference in means was due to chance
A p-value is the probability of observing a result as extreme as (or more extreme than) the one obtained, assuming the null hypothesis is true. P = 0.01 means there is a 1% probability the observed difference occurred by chance alone. It is not the probability of a Type II error (that is beta, or 1 - power), and it is not 0.1% (that would be p = 0.001).

24. What does the standard error of the mean (SEM) indicate?
Answer: The amount of error in the prediction of the population mean
The SEM estimates how precisely the sample mean reflects the true population mean. It is calculated as SD/√n. A smaller SEM means the sample mean is a more reliable estimate of the population mean. It is distinct from standard deviation (which describes variability of individual data points around the mean - that was the answer to Q5).

25. Factor most likely to lead to downgrading of an RCT's quality rating
Answer: There was evidence of inadequate concealment of treatment allocation in a double-blinded trial
Allocation concealment is a core methodological safeguard against selection bias. If investigators can predict which treatment a participant will receive before randomisation, they may (consciously or not) select patients accordingly, introducing bias that fundamentally undermines the RCT's internal validity. This is a direct threat to the integrity of randomisation and is a standard GRADE downgrading criterion. The other options (surrogate endpoints, single ethnicity, early termination, protocol amendments) can also lead to downgrading but are generally considered less fundamental than compromised allocation concealment.

26. Incidence and prevalence of breast cancer
Answer: Incidence 0.001/year and prevalence 0.05
  • Incidence = new cases per year / total population = 10 / 10,000 = 0.001 per year
  • Prevalence = total existing cases / total population = 500 / 10,000 = 0.05

27. Method to control confounding
Answer: Randomisation
Randomisation distributes both known and unknown confounders equally between study groups, making it the most effective method to control confounding at the design stage. Blinding reduces measurement/observer bias but does not control confounding. Matching and stratification are other valid approaches, but they are not listed here.

29. Type of study where children are randomly assigned to a new vs. existing vaccine and followed for safety and efficacy
Answer: Clinical trial
This is a randomised controlled trial (a type of clinical trial). Key features: random assignment to intervention groups, comparison of a new vs. existing treatment, prospective follow-up for both efficacy and safety outcomes. A prospective cohort study does not involve randomisation. The question specifically says "randomly assigned," which distinguishes this from an observational cohort design.

30. Next step in the audit cycle for the GP considering an amitriptyline audit
Answer: Identify standards
The GP has recognised a topic of interest (the needs assessment stage is effectively complete - they have identified a potential gap in practice). The next formal step in the audit cycle is to identify standards (i.e., what the evidence-based or guideline-recommended criteria for safe prescribing of amitriptyline actually are), before data can be collected and compared against those standards.
The audit cycle order is: Needs assessment → Identify standards → Data collection → Data analysis → Implement change → Re-audit
This is a shared conversation. Sign in to Orris to start your own chat.