Evidence-based practice1. A 52-year-old man, Tomasz Kowalczyk, is randomised to a 12-week weight-management programme. Eighteen percent of the intervention arm never attend. Analysts still include every randomised participant in the original groups. What does this analysis primarily estimate?A. The effect of offering the programme as randomised, including people who never attended a sessionB. The biological effect only in people who attended every session, which some claim is the sole unbiased estimate of efficacyC. Whether the randomisation sequence was generated with undisclosed block sizes of four throughout recruitmentD. The accuracy of a diagnostic index test against an agreed reference standard in the same sampleE. How much publication bias is present when small trials are missing from a funnel plot of effect versus precisionCheck
Evidence-based practice2. A 41-year-old woman, Sian Pritchard, asks what ‘specificity’ means after reading a paper on a practice natriuretic-peptide cut-off. Which statement is correct?A. The chance she personally has heart failure if this particular test comes back positive in clinicB. The proportion of people without the target condition who test negative on that cut-offC. The proportion of people who truly have the condition who also test positive on the same cut-offD. How much the odds of disease rise after a positive test, which is the definition of LR+E. The number of similar people you need to treat to prevent one extra clinical eventCheck
Evidence-based practice3. A journal club reviews a cluster trial of a new asthma template. The relative risk of exacerbation is 0.94 with a 95% confidence interval of 0.71 to 1.22. What is the most appropriate interpretation?A. The result proves harm because the upper bound of the interval exceeds oneB. The p-value must be below 0.001 simply because the point estimate sits below oneC. The interval includes 1, so a statistically significant relative effect is not shown at 5%D. Confidence intervals that overlap 1 can be ignored whenever a forest-plot diamond appears to sit left of the line of no effectE. NNT for counselling can be quoted directly as one divided by the relative risk 0.94Check
Evidence-based practice4. A 67-year-old man, Idris Bennett, has a positive faecal immunochemical test used in a low-prevalence surveillance clinic. The same assay had a PPV of 28% in a two-week-wait cohort. Which change is most likely in this lower-prevalence setting if sensitivity and specificity are unchanged?A. PPV will fallB. Sensitivity will automatically fall even if the laboratory cut-off and 2×2 definitions stay identicalC. NPV will fall in lockstep with PPV whenever a clinic moves from high to low prevalenceD. The positive likelihood ratio must become exactly 1 because prevalence has changedE. The confidence interval for any associated odds ratio is forced to include 1Check