Advanced journal club facilitation — CASC/teaching station · Psych CASC / OSCE

Prompt

You are the psychiatry registrar leading a 10-minute advanced journal club with a junior doctor, pharmacist, and nurse practitioner. The paper is a random-effects meta-analysis of antidepressants with high I-squared, an asymmetric funnel plot, and a GRADE summary rating moderate-to-low certainty. A consultant has asked whether the team should switch the ward default SSRI based on a network ranking figure in the discussion section.

Station brief

Format. Teaching/communication station, approximately 8–10 minutes. Facilitate, do not monologue. Check understanding. [4]

Candidate instructions. Help the team interpret the meta-analysis at an advanced but accessible level: forest plot and prediction interval, heterogeneity, funnel asymmetry, GRADE certainty vs recommendation strength, and why an NMA ranking should not automatically rewrite ward defaults. Close with a practical team action. [1][5][6]

Candidate scenario

Station materials (summary): Meta-analysis of SSRI X vs placebo, 11 trials, random-effects RR for response 0.82 (95% CI 0.74–0.91), I-squared 65%, prediction interval crosses 1, funnel asymmetric. Authors discuss a multipage NMA ranking table favouring drug Y as 'best'. Ward currently uses sertraline as default with monitoring protocols already embedded. Junior doctor asks: "So should we switch everyone to drug Y tomorrow?" [1][2][8]

Marking domains

Structures discussion: validity → results (including PI/heterogeneity) → publication-bias caution → GRADE → applicability/action [1][4]
Explains diamond vs prediction interval in plain language [1]
Explains I-squared as inconsistency proportion, not a magic stop rule [2]
Treats funnel asymmetry as small-study-effect signal needing caution (not automatic discard) [3][7]
Separates certainty of evidence from recommendation strength [4][5]
Explains NMA rankings need transitivity/absolute effects; not automatic formulary switch [6][8]
Collaborative tone; invites pharmacist on harms/interactions and nurse on monitoring workload [5]
Time management; clear bottom line [4]

Reveal assessor key

Open (30–45 s). "We'll use four steps: can we trust it, what do the numbers mean including inconsistency, how sure are we, and what should we do for our patients." Write PICO on the board. [4]

Validity/results. Point to the forest plot: average benefit (diamond) may be real, but the prediction interval crossing null means a new ward/setting might not see the same benefit — so we should not treat the diamond as destiny.[1] High I-squared means studies disagree more than chance alone; ask what differed (dose, severity, blinding).[2]

Funnel / publication bias. "The funnel looks lopsided. That can mean missing negative studies, true differences in small trials, or chance — psychiatry has a track record of selective publication of antidepressant trials, so we stay sceptical."[3][7]

GRADE. "Certainty is not high — risk of bias plus inconsistency plus publication concern. Even if we lean toward some average benefit, a strong blanket recommendation to switch defaults does not follow; recommendation strength also depends on values, harms, and systems."[4][5]

NMA ranking. "League tables mix direct and indirect comparisons. We need to know if the network is fair (transitivity) and what the absolute differences are. Rankings alone are a weak reason to overturn a working default with established monitoring."[6][8]

Close. Practical action: no immediate mass switch; pharmacist to compare metabolic/interaction profiles of Y vs current default; next journal club to walk RoB 2 on the two largest trials; document shared decision-making for individual switches. Thank the team. [5][6]

Common pitfalls

Lecturing without checking understanding; equating statistical significance with mandatory protocol change; ignoring prediction intervals; treating rankings as gold medals; dismissing the whole paper because of funnel asymmetry; failing to involve the multidisciplinary team on harms and workflow. [1][3][6]

References

[1]Riley RD, Higgins JPT, Deeks JJ Interpretation of random effects meta-analyses BMJ, 2011.PMID 21310794
[2]Higgins JP, Thompson SG, Deeks JJ, Altman DG Measuring inconsistency in meta-analyses BMJ, 2003.PMID 12958120
[3]Sterne JAC, Sutton AJ, Ioannidis JPA, et al. Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised controlled trials BMJ, 2011.PMID 21784880
[4]Guyatt GH, Oxman AD, Vist GE, et al; GRADE Working Group GRADE: an emerging consensus on rating quality of evidence and strength of recommendations BMJ, 2008.PMID 18436948
[5]Guyatt GH, Oxman AD, Kunz R, et al. Going from evidence to recommendations BMJ, 2008.PMID 18467413
[6]Mills EJ, Ioannidis JPA, Thorlund K, Schünemann HJ, Puhan MA, Guyatt GH How to use an article reporting a multiple treatment comparison meta-analysis JAMA, 2012.PMID 23011714
[7]Turner EH, Matthews AM, Linardatos E, Tell RA, Rosenthal R Selective publication of antidepressant trials and its influence on apparent efficacy N Engl J Med, 2008.PMID 18199864
[8]Cipriani A, Furukawa TA, Salanti G, et al. Comparative efficacy and acceptability of 21 antidepressant drugs for the acute treatment of adults with major depressive disorder: a systematic review and network meta-analysis Lancet, 2018.PMID 29477251

Station brief

Format. Teaching/communication station, approximately 8–10 minutes. Facilitate, do not monologue. Check understanding. [4]

Candidate scenario

Marking domains

Structures discussion: validity → results (including PI/heterogeneity) → publication-bias caution → GRADE → applicability/action [1][4]

Explains diamond vs prediction interval in plain language [1]

Explains I-squared as inconsistency proportion, not a magic stop rule [2]

Treats funnel asymmetry as small-study-effect signal needing caution (not automatic discard) [3][7]

Separates certainty of evidence from recommendation strength [4][5]

Explains NMA rankings need transitivity/absolute effects; not automatic formulary switch [6][8]

Collaborative tone; invites pharmacist on harms/interactions and nurse on monitoring workload [5]

Time management; clear bottom line [4]

Reveal assessor key

Open (30–45 s). "We'll use four steps: can we trust it, what do the numbers mean including inconsistency, how sure are we, and what should we do for our patients." Write PICO on the board. [4]

Common pitfalls

References

[1]Riley RD, Higgins JPT, Deeks JJ Interpretation of random effects meta-analyses BMJ, 2011.PMID 21310794
[2]Higgins JP, Thompson SG, Deeks JJ, Altman DG Measuring inconsistency in meta-analyses BMJ, 2003.PMID 12958120
[3]Sterne JAC, Sutton AJ, Ioannidis JPA, et al. Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised controlled trials BMJ, 2011.PMID 21784880
[4]Guyatt GH, Oxman AD, Vist GE, et al; GRADE Working Group GRADE: an emerging consensus on rating quality of evidence and strength of recommendations BMJ, 2008.PMID 18436948
[5]Guyatt GH, Oxman AD, Kunz R, et al. Going from evidence to recommendations BMJ, 2008.PMID 18467413
[6]Mills EJ, Ioannidis JPA, Thorlund K, Schünemann HJ, Puhan MA, Guyatt GH How to use an article reporting a multiple treatment comparison meta-analysis JAMA, 2012.PMID 23011714
[7]Turner EH, Matthews AM, Linardatos E, Tell RA, Rosenthal R Selective publication of antidepressant trials and its influence on apparent efficacy N Engl J Med, 2008.PMID 18199864
[8]Cipriani A, Furukawa TA, Salanti G, et al. Comparative efficacy and acceptability of 21 antidepressant drugs for the acute treatment of adults with major depressive disorder: a systematic review and network meta-analysis Lancet, 2018.PMID 29477251