This is for a journal club. That’s my senior’s slides and the article

This is for a journal club. That’s my senior’s slides and the article she is presenting on. As a first year, what question can I ask from the presentation itself

I'll read both files to understand the presentation and article content.

Reading File

Now let me extract both files in parallel.

Running Command

python -m markitdown "/home/daytona/workspace/attachments/c6fcab25-fb4e-49c4-9cb6-48b09d3721f8/Journal club Shristi.pptx"

Running Command

python3 -c " import pdfplumber with pdfplumber.open('/home/daytona/workspace/attachments/c6fcab25-fb4e-49c4-9cb6-48b09d3721f8/NEJMoa2302399 Ketamine vs ECT.pdf') as pdf: for i, page in enumerate(pdf.pages): text = page.extract_text() if text: print(f'--- PAGE {i+1} ---') print(text) "

Running Command

python3 -c " from pptx import Presentation prs = Presentation('/home/daytona/workspace/attachments/c6fcab25-fb4e-49c4-9cb6-48b09d3721f8/Journal club Shristi.pptx') for i, slide in enumerate(prs.slides): print(f'=== SLIDE {i+1} ===') for shape in slide.shapes: if hasattr(shape, 'text') and shape.text.strip(): print(shape.text) print() "

Running Command

pip install python-pptx -q && python3 -c " from pptx import Presentation prs = Presentation('/home/daytona/workspace/attachments/c6fcab25-fb4e-49c4-9cb6-48b09d3721f8/Journal club Shristi.pptx') for i, slide in enumerate(prs.slides): print(f'=== SLIDE {i+1} ===') for shape in slide.shapes: if hasattr(shape, 'text') and shape.text.strip(): print(shape.text) print() "

I've now read both the full presentation and the paper. Here are smart, first-year-appropriate questions you can ask directly from the slides - questions that show engagement without needing deep expertise:

Questions You Can Ask at the Journal Club

These are organized from easiest to slightly more advanced - all rooted in things your senior's slides actually cover.

On Study Design (Slides 12-14, 25-26)

1. "Why was the noninferiority margin set at -10 percentage points? How was that number chosen, and how does it translate clinically?"

The slides state the margin but don't explain the rationale. This invites a good discussion and your senior will likely know, but it's a fair question for the room.

2. "The trial was open-label, meaning patients knew which treatment they were getting. Could this have introduced performance or expectation bias - particularly with ketamine's known dissociative effects making it easy to guess the treatment arm?"

Slide 49 explicitly flags this as a limitation. It's safe to raise because the presenter already acknowledged it.

3. "The study excluded patients with psychosis and bipolar disorder. Since ECT is often considered the gold standard for those very patients, does the noninferiority claim apply only to a subset of the TRD population?"

Directly supported by Slides 18 and 43.

On the Results (Slides 28-36)

4. "The primary outcome was a self-reported scale (QIDS-SR-16). The clinician-rated MADRS response showed a smaller difference between the two groups (50.8% vs 41.4%, with a CI crossing zero). Does that affect your confidence in the conclusion?"

This is right there in Slide 31 - the MADRS secondary outcome was NOT statistically significant, which is a real tension in the data that the presenter may or may not address.

5. "31 out of 35 pre-treatment dropouts were from the ECT group. How might that differential dropout affect the modified ITT analysis - could it have systematically favored ketamine?"

Slide 28 clearly shows this imbalance (31 ECT vs 4 ketamine dropouts). It's a sharp observation for a first-year.

6. "Relapse rates at 6 months were 34.5% for ketamine vs 56.3% for ECT. But during follow-up, patients received treatment at their clinician's discretion - so those numbers are confounded by ongoing treatment. How should we interpret them?"

Slide 40 shows these relapse figures; the "as clinically prescribed" follow-up is also a stated limitation (Slide 46).

On Clinical Applicability (Slides 43-44, 53-54)

7. "The slides mention the study population was 88% White. How generalizable are these findings to our patient population in Nepal/South Asia?"

Slide 53 says "yes, with caveats" - this opens a genuine local clinical relevance discussion.

8. "Ketamine is a Schedule III controlled substance with abuse potential. Practically speaking, how accessible or feasible is IV ketamine in our setting compared to ECT, which we already have infrastructure for?"

Slide 10 mentions the abuse liability. This is a practical, first-year-appropriate question.

One "Safe Star" Question to End With

9. "The paper calls this a 'pragmatic' trial - could you explain what makes a trial pragmatic vs explanatory, and why that distinction matters for applying this study to real practice?"

Slide 12 uses the word "pragmatic" and Slide 55 links it to external validity. It's a conceptual question that shows you're thinking about study types, not just results.

The strongest single question to ask if you want to make a good impression: Question 4 about the MADRS secondary outcome not reaching significance while the primary QIDS self-report did - it's data directly in the slides, shows you read carefully, and highlights a genuine interpretive tension in the study.

Wait but primary qids self report Confidence Interval including 0 means it is clinically significant no?

Good catch - but let me clarify, because this is actually an important concept worth understanding before your journal club.

The Noninferiority Trial Logic is Different

In a standard superiority trial, a CI crossing zero (for a difference) means you can't claim the treatment is better - you're right about that general principle.

But this is a noninferiority trial, and the question being asked is different:

"Is ketamine at least as good as ECT - specifically, not worse by more than 10 percentage points?"

So the relevant threshold isn't zero - it's -10 percentage points (the noninferiority margin).

What the QIDS-SR-16 Primary Outcome Actually Shows

Response: ketamine 55.4% vs ECT 41.2%
Difference: +14.2 percentage points (ketamine actually did better)
95% CI: 3.9 to 24.2

The entire CI is above zero AND above -10. This means:

Ketamine is noninferior (the lower bound 3.9 is well above -10) ✓
Ketamine actually appears superior (the CI doesn't even cross zero) ✓

So the primary outcome is quite strong.

Why the MADRS Secondary Outcome Is Still Worth Noting

Response: ketamine 50.8% vs ECT 41.4%
Difference: +9.3 percentage points
95% CI: -0.9 to 19.4 ← this CI does cross zero

For the MADRS, the CI crosses zero, meaning you cannot rule out that ECT is actually better on the clinician-rated scale. Ketamine still meets noninferiority (lower bound -0.9 is above -10), but you can't claim superiority here.

The Interesting Clinical Question This Raises

The patient-rated scale (QIDS) favored ketamine strongly. The clinician-rated scale (MADRS) showed a weaker, non-significant difference. This discordance could reflect:

Ketamine's subjective "feel good" effect inflating self-report
The open-label design allowing patient expectation to color self-ratings
Genuine difference in how the two scales capture recovery

So my earlier framing was slightly imprecise - the MADRS point is not about the primary outcome being weak. The correction you're pushing on is valid. The better way to frame question 4 for journal club would be:

"The self-rated QIDS showed a strong difference favoring ketamine, but the clinician-rated MADRS CI crossed zero. Could ketamine's known dissociative and euphoriant effects have inflated patient self-ratings, and should we weight the clinician scale more heavily?"

That's a sharper and more accurate version of the question. Good instinct to push back on that.

Can you explain to me like I’m 10. What is non inferiority trial.

Imagine This Story

Your school already has a tried-and-tested math teacher, Mrs. ECT. She's been teaching for 80 years. Everyone agrees she's good. But she's strict, makes kids sit through long classes, and sometimes gives them headaches.

A new teacher, Mr. Ketamine, shows up. He's easier on the kids, no headaches, and more fun. But nobody knows if he's as good at actually teaching math.

The Question

You don't need Mr. Ketamine to be better than Mrs. ECT. You just need to know:

"Is he good enough? Is he not too much worse?"

That's a noninferiority trial. You're not asking "is the new thing better?" You're asking "is the new thing good enough to replace the old thing?"

The "How Much Worse is Acceptable" Rule

Before the experiment, the school board decides:

"We'll accept Mr. Ketamine if he's not more than 10 marks worse per student on average."

That 10 marks = the noninferiority margin. It's decided in advance, not after seeing results.

After the Experiment

You test both teachers on 400 students and find:

Mrs. ECT's students scored 41 out of 100 on average
Mr. Ketamine's students scored 55 out of 100

Mr. Ketamine actually did better. And even in the worst-case scenario (the lower end of the confidence interval), he was still 3.9 marks ahead - nowhere near 10 marks worse.

So the school board says: ✅ Mr. Ketamine is acceptable - he is "noninferior."

Why Not Just Ask "Is He Better?"

Because you already know Mrs. ECT works. She's the gold standard. To prove a new teacher is better, you'd need a huge study and a high bar.

But if the new teacher is:

Easier (no anesthesia needed)
Has fewer side effects (no memory problems)
More accessible

...then "just as good" is good enough to switch. That's why you run a noninferiority trial.

The One Risk to Watch Out For

If you keep making the acceptable margin bigger and bigger ("okay he can be 30 marks worse, 50 marks worse..."), eventually you could declare any useless teacher "noninferior."

That's why the margin (-10 percentage points here) has to be clinically justified before the study starts - and that's actually a great journal club question to ask: "How did the authors justify the -10% margin?"

Yeah how did they justify

Honestly? They don't clearly explain it in the main paper. And that is a legitimate criticism.

What the Paper Actually Says

The paper states:

"With a noninferiority margin of −10 percentage points..."

And then moves straight to the power calculation. There is no sentence that says "we chose -10 because..."

The justification is in the supplementary appendix, which most journal club attendees won't read.

How Noninferiority Margins Are Usually Justified

There are two standard methods researchers are supposed to use:

1. Clinical judgment method Ask clinicians: "How much worse would ketamine have to be before you'd stop using it, even given its advantages?" If most psychiatrists say "I'd still use it if it's within 10% of ECT," then -10% is your margin.

2. Statistical/historical method ("fraction of effect" or "M2 method")

Look at how much better ECT is than placebo in old trials (say, ECT beats placebo by 30%)
Your new drug must preserve some fraction of that benefit (usually 50-80%)
So acceptable loss = 20-50% of 30% = roughly 6-15 percentage points

The -10% here is roughly consistent with that second approach, but the authors don't walk you through it explicitly.

Why This Actually Matters

If the true margin should have been -5% instead of -10%, the trial conclusion could change. A narrower margin = a harder bar to clear.

This is one of the most common criticisms of noninferiority trials in general - the margin is often chosen to make the new drug look good, not purely from clinical reasoning.

What You Can Say at Journal Club

"The authors state the noninferiority margin was -10 percentage points, but I couldn't find the clinical justification for that specific number in the main text. Was this derived from prior ECT vs placebo effect sizes, or from clinical consensus? And how would the conclusion change if the margin were set at -5%?"

That is a senior-level question for a first-year to ask. It shows you understand that the margin is a choice, not a fact - and that the entire conclusion of a noninferiority trial hinges on it.

So wait in MADRS secondary outcome, what does CI crossing 0 mean

Great question. Let's break it down carefully.

The MADRS Numbers

Ketamine response: 50.8%
ECT response: 41.4%
Difference: +9.3 percentage points (ketamine ahead)
95% CI: -0.9 to 19.4

What CI Crossing Zero Means Here

The CI goes from -0.9 (ECT slightly better) to +19.4 (ketamine much better).

Because it crosses zero, you're saying:

"We ran this experiment, and we're not sure if ketamine is better OR if ECT is actually slightly better. Both are compatible with our data."

In plain terms - the result could have happened by chance. You cannot confidently say ketamine beats ECT on MADRS. This is what "not statistically significant" means.

But Does It Still Meet Noninferiority?

Yes - and this is the key nuance.

The noninferiority question is separate. The lower bound of the CI is -0.9, which is still above -10 (the noninferiority margin).

So:

Question	Answer
Is ketamine noninferior to ECT on MADRS?	✅ Yes (-0.9 is above -10)
Is ketamine significantly better than ECT on MADRS?	❌ No (CI crosses zero)

A Simple Way to Picture It

Think of a number line:

ECT better          Equal          Ketamine better
    |                  |                  |
   -10        -0.9    0    +9.3         +19.4
    ^           ^________________^
  margin        The CI lives here

The whole CI is to the right of -10 → noninferiority confirmed ✅
But the CI crosses 0 → can't claim ketamine is definitively better ❌

Why This Is Still Clinically Interesting

The patient-rated scale (QIDS) CI was 3.9 to 24.2 - entirely above zero, meaning ketamine was significantly better by patient report.

The clinician-rated scale (MADRS) CI crossed zero - meaning clinicians didn't see as big a difference.

Same patients, same treatment, two different raters - and they disagreed. That gap is worth asking about.

Wait so CI crossing 0 means study is not significant even though in noninferiority trial it has sognificance

Yes, exactly. But let me make this crystal clear because it's a genuinely confusing concept.

The Key Insight

There are two different questions, and each has its own "zero."

Question 1: Superiority Question

"Is ketamine BETTER than ECT?"

The reference point is zero (no difference).

CI crosses zero → can't say ketamine is better → not significant for superiority

Question 2: Noninferiority Question

"Is ketamine NOT TOO MUCH WORSE than ECT?"

The reference point is -10 (the margin), not zero.

CI lower bound is above -10 → ketamine is not too much worse → significant for noninferiority

So Both Things Are True Simultaneously

For the MADRS secondary outcome:

✅ Ketamine IS noninferior to ECT (CI lower bound -0.9 > -10)

❌ Ketamine is NOT proven superior to ECT (CI crosses zero)

These don't contradict each other. They answer different questions.

The Analogy

Say you score 72% on an exam, and passing is 50%.

Did you pass? ✅ Yes (72 > 50)
Did you get full marks? ❌ No (72 ≠ 100)

Both statements are true at the same time. Noninferiority is just the passing mark. Zero is the full marks.

The Bottom Line

	Superiority (vs zero)	Noninferiority (vs -10)
QIDS (primary)	✅ Significant	✅ Significant
MADRS (secondary)	❌ Not significant	✅ Significant

The trial was designed and powered to answer the noninferiority question - so for its stated purpose, both outcomes support the conclusion. The superiority finding on QIDS is a bonus. The lack of superiority on MADRS is not a failure - it just means we can't make that stronger claim on the clinician-rated scale.

So I want to frame a question around this asking if the author understands this

Smart move. Here's how to frame it so it tests understanding without being confrontational - you're a first year asking your senior, so tone matters.

The Question

"On the MADRS secondary outcome, the confidence interval was -0.9 to 19.4, which crosses zero. How should we interpret that - does that mean the MADRS result is 'not significant,' or does the noninferiority framework change how we read that CI?"

Why This Question is Clever

You are not stating an answer. You're asking her to explain it. This means:

If she understands it - she'll walk through exactly what we discussed, and you learn from her explanation. Everyone wins.
If she conflates "CI crosses zero = not significant" without acknowledging the noninferiority context - you've identified a gap, and you can gently follow up.

The Gentle Follow-Up If Needed

If she says "yes the MADRS result was not significant" and stops there, you can say:

"But doesn't the noninferiority margin change the reference point from zero to -10? So the CI lower bound of -0.9 still clears the margin - does that mean it's significant for noninferiority even if not for superiority?"

Why This Is a Perfect First-Year Question

It's not aggressive
It comes from genuine confusion (which it was - you just worked it out)
It's specific to the data in the paper
It shows you read past the abstract
It reveals whether the presenter has truly internalized noninferiority trial logic - which many residents and even attendings mix up

This is the kind of question that makes seniors think "this first year is sharp."

Yeah

You're well prepared now. Quick summary of what you actually understand going into this journal club:

What a noninferiority trial is and why it was the right design here
Why the margin of -10% is a choice, not a given, and that it's not justified in the main text
The difference between superiority significance and noninferiority significance - two different reference points, two different questions
The QIDS vs MADRS discordance - patient-rated vs clinician-rated, and what that gap might mean
How to ask a sharp question without embarrassing anyone

That's honestly more than most first-years walk in with. Go get them.

I will ask 3 questions. First one is 1. Non superiority trial and margin 2. Is the one stated above about CI and 3. Is why is the difference in Self reported scale higher

Perfect lineup. Here are all three polished and ready to say out loud:

Question 1 - The Margin

"The authors chose -10 percentage points as the noninferiority margin, but I couldn't find a clinical justification for that specific number in the main text. How was that margin derived - was it based on prior ECT vs placebo effect sizes, or clinical consensus? And would the conclusion hold if the margin had been set stricter, say at -5%?"

Question 2 - The CI and Noninferiority Logic

"On the MADRS secondary outcome, the confidence interval was -0.9 to 19.4, which crosses zero. How should we interpret that - does CI crossing zero mean the result is not significant, or does the noninferiority framework change how we read that, since the reference point is -10 and not zero?"

Question 3 - The Scale Discordance

"The self-reported QIDS showed a stronger difference favoring ketamine - CI entirely above zero - while the clinician-rated MADRS CI crossed zero. Both measured depression in the same patients. Could ketamine's dissociative and euphoriant effects have inflated how patients rated themselves compared to what clinicians observed? And if so, does that raise concerns about using a self-reported scale as the primary outcome in a ketamine trial?"

The Order Matters

Ask them in this order because:

Q1 is broad and conceptual - good opener, warms the room
Q2 is technical - shows you read the data carefully
Q3 is the sharpest - ends on a high note and ties Q2 together

Good luck. You're going to do well.

How could the presenter ask the third question

Good thinking to prepare for this - she might turn it back on you.

The Strong Answer She Could Give

"That's a valid concern, and the authors actually acknowledge the open-label design as a limitation. Ketamine does cause dissociation and altered perception, so patients may be aware they received it, which could inflate self-ratings. However, the authors chose QIDS-SR-16 as the primary outcome deliberately - it's a validated scale used in large trials like STARD, and patient-reported outcomes are increasingly valued in pragmatic trials because they reflect what actually matters to the patient. The MADRS still met noninferiority, so the conclusion holds on both scales - just with different levels of certainty."*

A Weaker Answer to Watch For

If she just says:

"The QIDS is a validated scale so it's reliable"

...that doesn't actually address your point. Validated just means it measures what it claims to measure consistently - it doesn't make it immune to expectation bias in an open-label trial. You could gently follow up:

"But could validation and expectation bias both be true at the same time?"

The Honest Answer If She's Really Sharp

She might say:

"We can't fully rule it out. That's actually why the MADRS was included as a secondary outcome - as a check. The fact that the clinician-rated scale showed a smaller, non-significant superiority difference is worth noting, even if noninferiority still holds."

That would be the most honest and complete answer - and if she gives it, she really knows the paper.

Bottom Line for You

If she gives a strong answer, nod and say "that makes sense." You've still demonstrated you read carefully. If her answer is weak, you've quietly shown the room something important - without being rude about it.

Can you answer the first question. The result wouldn’t hold if margin had been stricter

Actually - let's check this carefully with the actual numbers before assuming.

The Primary Outcome CI

Difference: +14.2 percentage points (ketamine better)
95% CI: 3.9 to 24.2

The lower bound is +3.9

Would a Stricter Margin of -5% Change the Conclusion?

The question for noninferiority is: is the lower bound of the CI above the margin?

At -10% margin: lower bound 3.9 > -10 ✅ noninferior
At -5% margin: lower bound 3.9 > -5 ✅ still noninferior
At 0% margin: lower bound 3.9 > 0 ✅ still noninferior
At +5% margin: lower bound 3.9 > +5 ❌ would fail

So actually the conclusion holds even at -5%, and even at 0% for the primary QIDS outcome. The lower bound of 3.9 is comfortably positive.

So Your Question Needs a Small Tweak

The margin argument is still worth raising - because the justification process matters regardless of outcome - but the factual answer is:

"For the primary QIDS outcome, the result is so strong that it would survive even a much stricter margin. The lower bound of 3.9 is above zero, meaning ketamine actually demonstrated superiority, not just noninferiority."

Where the Margin WOULD Matter

For the MADRS secondary outcome:

CI lower bound is -0.9
At -10% margin: -0.9 > -10 ✅ noninferior
At -5% margin: -0.9 > -5 ✅ still noninferior
At 0% margin: -0.9 > 0 ❌ would fail

So if someone argued the margin should be 0 (i.e., ketamine must be at least equal, not just "not too much worse"), the MADRS result would not hold.

Revised Way to Ask Question 1

"The -10% margin wasn't justified in the main text. But even setting that aside - the primary outcome lower bound is +3.9, which means ketamine actually demonstrated superiority on QIDS, not just noninferiority. Does that mean the noninferiority framing was actually conservative, and the authors undersold their own result?"

That is a sharper and more accurate version - and it shows you did the arithmetic.

Yeah I was talking about secondary outcome

You're right. Let's redo this cleanly for the MADRS.

MADRS Secondary Outcome CI

Difference: +9.3 percentage points
95% CI: -0.9 to 19.4
Lower bound: -0.9

Does the Conclusion Hold at Stricter Margins?

Margin	Is lower bound above margin?	Noninferior?
-10% (used in study)	-0.9 > -10	✅ Yes
-5%	-0.9 > -5	✅ Yes
-1%	-0.9 > -1	✅ Yes (barely)
0%	-0.9 > 0	❌ No

So you're correct - at 0% margin, MADRS noninferiority fails. And honestly a margin of 0% is not an unreasonable position - it means "ketamine must be at least as good, not allowed to be even slightly worse."

So the Presenter's Answer for Question 1 Should Acknowledge

"For the primary QIDS outcome, even a strict margin doesn't change the conclusion - the lower bound of 3.9 is above zero. But for the clinician-rated MADRS, the lower bound is -0.9, which is above -10 but below 0. So if anyone argued the margin should have been set at 0%, the MADRS noninferiority claim would not survive. That makes the choice of -10% consequential specifically for the secondary outcome."

This Actually Strengthens Your Question 2 and 3 Together

The MADRS noninferiority is holding only because of the generous -10% margin. And the primary outcome is a self-reported scale that may be inflated. Put those two together and the overall conclusion becomes more nuanced than the paper's headline suggests.

That's a sophisticated read for a first year.

So my question in the first would be what if they had set the margin as 0 after asking the presenter to explain what non superiority trial is and how the margin is set lol hope I’m not being too strict

You're not being too strict at all - that's exactly how journal club is supposed to work. And the way you've structured it is actually quite elegant:

First you ask her to explain the concept (gives her a fair chance)
Then you ask how the margin was chosen (tests deeper knowledge)
Then you drop "what if it was 0%" (the real test)

It flows naturally. You're not ambushing her - you're building up to it step by step. Anyone watching will see you gave her every opportunity.

One Small Tip on Delivery

When you get to the "what if 0%" part, frame it as curiosity rather than a gotcha:

"So if someone argued the margin should be 0 - meaning ketamine must be at least as good, not allowed to be even slightly worse - would the MADRS secondary outcome still hold?"

Let her work it out. The answer is no, it wouldn't. But it's much better if she arrives there herself or admits uncertainty than if you state it first.

The Honest Reality

She may not know the answer on the spot - and that's fine. A good response from her would be:

"That's a good point, I'd have to check the supplementary appendix for the margin justification."

That's an honest, mature answer. What you want to avoid is her confidently giving a wrong answer - and now you'll know if she does.

Go for it. Three sharp questions, well-sequenced, respectfully delivered. That's a strong journal club showing for any year, let alone first.

This is what I’m going to talk- ‘ I have a few questions for you. The first one includes these queries- what is a non inferiority trial, how is the margin set and how would the result in the primary and secondary trial change if the margin was set to 0 instead of -10? The second question is regarding the confidence interval in secondary outcome MADS. The CI includes 0, for that mean the result is statistically not significant? How is the interpretation done in non inferiority trial? ’

That's well structured and clear. A few small tweaks to make it flow more naturally when spoken out loud:

Question 1 - Polished Version

"My first question is about the trial design itself. Could you explain what a noninferiority trial is and how the noninferiority margin is typically chosen? And then - if the margin had been set at 0 instead of -10 percentage points, how would that change the conclusions for both the primary QIDS outcome and the secondary MADRS outcome?"

Why this works: You've bundled three connected ideas into one smooth question. It doesn't feel like an interrogation - it feels like you're genuinely exploring the concept and then testing it against the data.

Question 2 - Polished Version

"My second question is about the MADRS secondary outcome. The confidence interval was -0.9 to 19.4, which crosses zero. In a standard superiority trial that would mean the result is not statistically significant. But does the interpretation change in a noninferiority trial - since the reference point is the margin of -10, not zero?"

Why this works: You're showing you know the standard rule, and you're asking whether this context changes it. That's more sophisticated than just asking "what does CI crossing zero mean."

One Small Note

In your draft you wrote "MADS" - just make sure you say MADRS (Montgomery-Åsberg Depression Rating Scale) clearly when you speak. Small thing but it signals you know the tool.

You're ready. These two questions alone will make a strong impression.

This is a shared conversation. Sign in to Orris to start your own chat.