TRANSCRIÇÃO 🔗
WEBVTT
00:00:06.480 --> 00:00:12.880
Hello dear colleagues and welcome to another episode of Positive Psychiatry with Raket Jin.
00:00:13.439 --> 00:00:15.839
I hope things are going well for you.
00:00:16.399 --> 00:00:26.960
Listen, today I wanted to talk to you about the topic of how to become a better scientific reader in psychiatry.
00:00:28.239 --> 00:00:34.880
Well, the challenge for us in psychiatry is there are a lot of publications coming our way.
00:00:35.359 --> 00:00:38.159
But I think I should begin with a confession.
00:00:38.399 --> 00:00:50.719
And this confession comes after decades of personally reading psychiatric research, conducting psychiatric research, teaching psychopharmacology, participating in scientific discussions.
00:00:50.880 --> 00:01:00.079
I've become increasingly convinced that the greatest challenge facing us clinicians today isn't access to scientific information.
00:01:00.240 --> 00:01:02.079
That's not a problem anymore.
00:01:02.320 --> 00:01:08.159
It's knowing how to read it well, how to extract the greatest meaning from it.
00:01:08.560 --> 00:01:20.640
We are drowning in information, and every single week brings new randomized controlled trials and meta-analyses and systemic reviews and real-world evidence.
00:01:21.040 --> 00:01:23.599
Lots comes at us.
00:01:23.840 --> 00:01:30.079
And artificial intelligence, well, it's actually increased the number of scientific papers.
00:01:30.560 --> 00:01:33.200
But here's the uncomfortable question.
00:01:33.760 --> 00:01:38.159
Are we becoming better scientific readers and thinkers?
00:01:38.719 --> 00:01:43.680
Or are we simply consumers of scientific conclusions?
00:01:44.079 --> 00:01:47.359
I worry that it may sometimes be the latter.
00:01:48.079 --> 00:02:07.120
As I'm visiting colleagues, it becomes more and more clear that they are being spoon-fed information, and we must resist that temptation because scientific literacy requires more than understanding what the author's conclusions might be.
00:02:07.519 --> 00:02:11.680
It actually requires us to understand what the data actually means.
00:02:11.840 --> 00:02:13.120
What does it demonstrate?
00:02:13.280 --> 00:02:14.960
What does it not demonstrate?
00:02:15.039 --> 00:02:20.639
And how confidently we can translate those findings into clinical practice.
00:02:20.960 --> 00:02:27.439
So this is what I thought I would do for all of my colleagues who are interested in positive psychiatry.
00:02:27.599 --> 00:02:36.240
I have put together a list of 12 habits I believe can make us substantially better scientific readers.
00:02:36.479 --> 00:02:38.800
Shall we go on that journey together?
00:02:39.039 --> 00:02:40.560
So let's do that.
00:02:40.800 --> 00:02:45.439
So the first thing is let's call it habit number one.
00:02:46.240 --> 00:02:47.840
And this is what it is.
00:02:48.159 --> 00:02:56.960
I recommend you read the abstract last, or at least read it again at the very end.
00:02:57.840 --> 00:03:02.400
I know, I know that sounds almost, well, heretical.
00:03:02.800 --> 00:03:10.479
The abstract is where most of us begin, and unfortunately, where many of us end, me included.
00:03:10.879 --> 00:03:22.000
But please remember, abstracts are carefully compressed narratives, and they are the narratives based on what the writers want us to believe.
00:03:22.159 --> 00:03:26.879
They reflect what the investigators believe is the most important thing about their findings.
00:03:26.960 --> 00:03:33.599
I'm not at all saying that they exaggerate or they hide, but that's just human nature.
00:03:33.759 --> 00:03:35.280
That's how it goes.
00:03:35.599 --> 00:03:40.240
But that isn't necessarily the same as what is most clinically important.
00:03:40.719 --> 00:03:42.719
So here's my recommendation.
00:03:43.120 --> 00:03:49.439
Begin reading a paper with the research question in mind first.
00:03:50.319 --> 00:03:57.759
Then examine the methods, the primary outcome, the actual results, and the limitations.
00:03:58.479 --> 00:04:00.960
Then return to the abstract.
00:04:01.280 --> 00:04:07.680
Because if you don't do that, you will already have a preconceived notion of what the study showed.
00:04:08.479 --> 00:04:18.639
So you may not want to start reading a paper by reading the abstract first, even though that is the natural human inclination.
00:04:19.120 --> 00:04:19.439
Okay?
00:04:20.399 --> 00:04:37.920
So give my habit number one recommendation some careful thought because you might see that there is wisdom in doing it the slightly harder way rather than the easy way, which is to read just what the authors want us to believe.
00:04:38.480 --> 00:04:40.560
Well, here's habit number two.
00:04:41.439 --> 00:04:45.439
Fall in love with the methods section.
00:04:46.319 --> 00:04:55.920
So if the abstract is like a movie trailer, the methods section of an article is where we discover how the movie was actually made.
00:04:56.240 --> 00:05:00.800
And in scientific research, methods matter enormously.
00:05:01.040 --> 00:05:09.600
So when reading a psychiatric clinical trial article or a paper, I ask myself and encourage you to do the same.
00:05:09.839 --> 00:05:11.360
Ask the following questions.
00:05:11.600 --> 00:05:12.879
Who was included?
00:05:13.120 --> 00:05:18.000
And perhaps as importantly, who was excluded from the study?
00:05:18.240 --> 00:05:21.920
Was the study adequately randomized and blinded?
00:05:22.399 --> 00:05:24.720
Was the comparator appropriate?
00:05:24.959 --> 00:05:26.800
Was there even a comparator?
00:05:27.439 --> 00:05:31.439
Was the trial long enough to answer the clinical question?
00:05:32.399 --> 00:05:37.519
Another set of questions to ask might be, how were the missing data handled?
00:05:38.160 --> 00:05:41.759
Was the primary outcome measure pre-specified?
00:05:42.240 --> 00:05:44.959
Was the study sufficiently powered?
00:05:45.680 --> 00:05:53.199
Now, some of these questions are relevant, but
Load full transcript
we've not been trained to think in this manner, have we?
00:05:53.439 --> 00:06:01.360
But I'm really encouraging you to do so if you want to become a high-quality reader of scientific articles.
00:06:01.920 --> 00:06:12.959
So consider maybe a depression trial that excludes patients with significant anxiety or substance use disorders or suicidal ideations or multiple medical comorbidities.
00:06:13.279 --> 00:06:22.639
So the findings from the paper may be scientifically valid, but it may not approximate the patients that you see in your waiting room.
00:06:23.199 --> 00:06:33.519
So internal validity tells us what we can trust in the study, but external validity helps us decide whether we can actually use it or not.
00:06:33.759 --> 00:06:35.360
And both matter.
00:06:36.000 --> 00:06:41.920
Now you see why falling in love with the methods section is so useful, right?
00:06:43.199 --> 00:06:45.120
Okay, habit number three.
00:06:45.439 --> 00:06:54.240
As a high-quality scientific reader, you should never confuse statistical significance with clinical significance.
00:06:54.720 --> 00:06:59.519
And this is a really important habit that we must all develop.
00:06:59.759 --> 00:07:10.000
Because a p-value below 0.05 tells us something about statistical compatibility with the null hypothesis.
00:07:10.560 --> 00:07:15.920
It does not tell us whether a treatment is clinically meaningful.
00:07:16.240 --> 00:07:20.160
It might be, but it doesn't automatically tell us that.
00:07:20.560 --> 00:07:30.240
So imagine a large antidepressant trial that demonstrates a statistically significant difference of 1.5 points on the madras scale.
00:07:30.399 --> 00:07:32.000
Should we celebrate this?
00:07:32.240 --> 00:07:33.439
Well, maybe.
00:07:33.759 --> 00:07:36.319
But we must ask additional questions.
00:07:36.480 --> 00:07:38.720
That's what a good scientific reader would do.
00:07:38.879 --> 00:07:41.759
They would ask, what was the baseline severity?
00:07:42.000 --> 00:07:43.519
What was the effect size?
00:07:43.759 --> 00:07:46.959
What proportion of patients achieved response or remission?
00:07:47.439 --> 00:07:49.839
What was the adverse event burden?
00:07:50.079 --> 00:07:52.319
How precise was this estimate?
00:07:52.879 --> 00:08:05.680
A statistically significant finding may in fact have large clinical importance, modest clinical importance, small clinical importance, or no clinical importance.
00:08:06.160 --> 00:08:11.519
You and I have to read the paper more carefully.
00:08:11.839 --> 00:08:15.920
I also want to tell you, sometimes the complete opposite can happen.
00:08:16.160 --> 00:08:27.199
We might have a p-value that is larger than 0.05, but the findings of the paper have very large clinical importance.
00:08:27.920 --> 00:08:35.600
So, habit number three educates us on the following: the p-value is a statistical verdict.
00:08:35.840 --> 00:08:38.399
It is not a clinical verdict.
00:08:38.639 --> 00:08:42.480
It is just one piece
of the evidence.
00:08:43.519 --> 00:08:44.399
Alright?
00:08:44.720 --> 00:08:55.200
Now we move on to habit number four, which is read the absolute numbers, not just the relative numbers.
00:08:55.519 --> 00:08:57.679
So I'll start with an example.
00:08:58.000 --> 00:09:02.639
Suppose a new intervention is reported to increase remission rates by 50%.
00:09:03.440 --> 00:09:06.000
Well that sounds impressive, doesn't it?
00:09:06.480 --> 00:09:13.519
But imagine remission increases from 20% to 30%, so the relative increase is 50%.
00:09:14.480 --> 00:09:19.200
However, the absolute increase is only 10% points.
00:09:19.440 --> 00:09:24.240
So this shows us the numbers needed to treat is in fact 10.
00:09:24.639 --> 00:09:28.799
So now we have information that can support a clinical conversation.
00:09:29.200 --> 00:09:35.440
We don't just want to read the absolute numbers, we want to read the relative numbers too.
00:09:35.919 --> 00:09:48.000
Because patients experience absolute benefits and harms, not relative numbers, not relative harms, not relative headlines.
00:09:48.240 --> 00:10:01.519
So please be careful as you read the paper because very often authors are more interested in reporting relative numbers and not absolute numbers, and it's our job to look for both of them.
00:10:02.320 --> 00:10:03.039
Very good.
00:10:03.200 --> 00:10:09.039
We move on to habit number five, which is my advice to do the following.
00:10:09.440 --> 00:10:15.120
Develop a healthy skepticism towards subgroup analyses.
00:10:15.600 --> 00:10:26.799
Now, subgroup analyses like studying just men or studying just women or studying people of a certain age or people with whatever other marker is important.
00:10:27.759 --> 00:10:41.759
But please do remember, unless it is pre-specified, sometimes it is a fishing trip by the authors or the sponsors, and that may or may not be the right way to go.
00:10:42.480 --> 00:10:53.279
So psychiatry is full of clinically meaningful subgroups, so patients with anxious depression or patients with early trauma or treatment resistant or inflammatory disorders.
00:10:53.440 --> 00:10:56.000
Yes, we do need to know about all of that.
00:10:56.240 --> 00:11:00.159
But I do urge caution unless it is pre-specified.
00:11:00.399 --> 00:11:05.440
Subgroup analyses require caution in making interpretations.
00:11:05.679 --> 00:11:11.519
And a good set of writers will actually alert you to that in the paper.
00:11:11.759 --> 00:11:23.120
But if they don't, you and I still should be quite thoughtful before we accept the conclusions about subgroup analysis as solid science.
00:11:23.360 --> 00:11:28.080
Because subgroup findings generate hypotheses.
00:11:28.639 --> 00:11:38.960
They do not automatically establish clinical truths unless it was pre-specified before the study is started.
00:11:39.759 --> 00:11:40.080
Okay?
00:11:41.279 --> 00:11:42.000
Very good.
00:11:42.159 --> 00:11:47.360
We're halfway, almost halfway through the journey, but let's move on to habit number six.
00:11:47.519 --> 00:11:55.279
And this says the following pay attention to what the study did not measure.
00:11:56.159 --> 00:11:57.919
What did the study not measure?
00:11:58.080 --> 00:11:59.200
We want to know that.
00:11:59.360 --> 00:12:05.120
And this is especially important in psychiatry because we frequently measure symptom reduction.
00:12:05.759 --> 00:12:11.279
But how often do we measure for restoration of functioning or quality of life?
00:12:11.519 --> 00:12:18.399
Or cognition, relationships, motivation, sleep, ability to experience pleasure.
00:12:18.960 --> 00:12:24.799
All the elements that are important to those of us interested in positive psychiatry.
00:12:25.679 --> 00:12:33.679
So a treatment might reduce a depression-rating scale score without producing a meaningful improvement in patients' daily life.
00:12:33.840 --> 00:12:35.840
That's a possibility, isn't it?
00:12:36.480 --> 00:12:46.799
So paying attention to what is not measured is the skill set of a high-quality scientific reader like yourself.
00:12:47.279 --> 00:12:53.679
So symptom reduction matters enormously, but it is not the entirety of recovery.
00:12:54.000 --> 00:13:00.320
If a paper actually measures recovery and wellness, we pay attention to it.
00:13:00.559 --> 00:13:04.960
But if it doesn't, we pay attention to that as well.
00:13:06.320 --> 00:13:12.080
All right, folks, now habit number seven, and we're going to ramp up a little bit.
00:13:12.320 --> 00:13:19.519
We're going to pressure test ourselves and encourage ourselves to learn about statistics a little bit.
00:13:19.840 --> 00:13:26.799
Because habit number seven says, learn to love confidence intervals.
00:13:27.120 --> 00:13:38.639
We'll talk about what it is in just a second, but confidence intervals are amongst the most underappreciated tools in scientific reading.
00:13:38.960 --> 00:13:43.279
Because when you say mean, that's just a point estimate.
00:13:43.440 --> 00:13:52.000
It tells us the observed effect when you pull every single participant and then you have one simple number.
00:13:52.240 --> 00:13:58.960
The problem is the simple number does not tell us what is the scatter around it.
00:13:59.360 --> 00:14:09.600
But a confidence interval helps us understand the precision of the mean and the range of effects that we find with the data.
00:14:09.840 --> 00:14:11.039
So do this.
00:14:11.679 --> 00:14:13.919
Let's look at an example, shall we?
00:14:14.159 --> 00:14:20.720
Consider two studies reporting the same estimated effect in terms of mean change.
00:14:20.960 --> 00:14:21.279
Okay?
00:14:21.679 --> 00:14:27.759
Now one study has a narrow confidence interval, but the other has a wide interval.
00:14:28.399 --> 00:14:31.200
Well, they're two different studies actually.
00:14:31.440 --> 00:14:43.759
They both had the same mean, but the one that had the narrow confidence interval will tell you that most people, most people in
that study got that benefit.
00:14:44.080 --> 00:14:56.720
If you see a very wide confidence interval, what you can then say, some people really, really got even more than that, but a lot of people got way less effect.
00:14:57.440 --> 00:15:06.799
So just because the mean on two different studies looks the same doesn't tell you very much about how the entire group behaved.
00:15:07.039 --> 00:15:13.279
So knowing about mean and knowing the confidence interval is very important.
00:15:14.159 --> 00:15:20.639
Most scientific papers report it, but generally speaking, we don't pay attention to it.
00:15:20.960 --> 00:15:28.480
But because you are interested in becoming a better scientific reader, I hope you won't do that anymore in the future.
00:15:28.799 --> 00:15:41.039
Keep your eyes on the confidence interval as much as you do as much as you do all other numbers when you're reading your paper.
00:15:42.240 --> 00:15:44.639
We move on to habit number eight.
00:15:44.879 --> 00:15:52.960
And this one says, be particularly careful with conclusions drawn from meta-analyses.
00:15:53.279 --> 00:15:55.919
Now, I have to be quite honest with you.
00:15:56.080 --> 00:16:04.639
I love a well-conducted meta-analysis, and nothing replaces it, because a meta-analysis is so important.
00:16:04.879 --> 00:16:11.360
It takes studies, multiple studies, with different drugs that sometimes cannot be done.
00:16:11.679 --> 00:16:18.879
No study can study 15 different medications, but meta-analysis, we can understand issues better.
00:16:19.120 --> 00:16:26.480
But our worry when we treat meta-analysis as the final word, it is definitely not.
00:16:26.799 --> 00:16:34.480
A meta-analysis is only as informative as the studies it includes and the assumptions underlying the synthesis.
00:16:34.799 --> 00:16:40.480
So look carefully when you read a meta-analysis at the heterogeneity of the studies.
00:16:40.960 --> 00:16:55.679
Well, examine the risk of bias, consider publication bias, selection bias, and whether the included studies evaluated comparable populations, where the intervention is kind of similar, were the comparative somewhat similar.
00:16:56.000 --> 00:17:03.840
Because if not, then the meta-analysis results may look impressive, but the conclusions might be a problem.
00:17:04.559 --> 00:17:11.839
So pooling studies does not automatically eliminate the weaknesses of individual studies, right?
00:17:12.480 --> 00:17:16.880
So absolutely read meta-analysis and psychiatry.
00:17:17.039 --> 00:17:30.880
Meta-analyses are incredibly important, but you may want to consider them as supportive, not the very final word on a particular scientific question.
00:17:31.200 --> 00:17:31.519
Okay?
00:17:32.400 --> 00:17:36.880
We move on to habit number nine, and it's a very important one.
00:17:37.839 --> 00:17:46.319
Let's learn to distinguish between evidence and interpretation of the evidence by the authors.
00:17:47.119 --> 00:17:53.119
Look, I've been an author before on multiple papers, and most likely so have you.
00:17:53.680 --> 00:18:00.240
There is an automatic bias we authors have to put our data in the best possible light.
00:18:00.400 --> 00:18:02.000
That's just human nature.
00:18:02.400 --> 00:18:08.720
But you and I, as scientific readers, have to distinguish between evidence and interpretation.
00:18:08.960 --> 00:18:15.680
And this habit requires intellectual discipline, because every scientific paper contains at least two stories.
00:18:16.000 --> 00:18:23.119
The first story is told by the data, but the second story is told by the authors.
00:18:23.359 --> 00:18:25.440
They're not always identical.
00:18:25.759 --> 00:18:28.720
So maybe we could think of an example.
00:18:29.039 --> 00:18:39.839
Consider a neuroimaging study that demonstrates an association between the activity in a particular brain neural network and depressive symptom severity.
00:18:40.319 --> 00:18:42.559
That finding is fascinating.
00:18:42.880 --> 00:18:52.000
But it does not establish causation, nor does it establish that manipulating that network will improve depression.
00:18:52.559 --> 00:19:05.680
So our job then is to look at the evidence, read the interpretation from the authors, but then decide: is it just exactly the right interpretation?
00:19:06.160 --> 00:19:11.119
Or maybe it is a slight exaggeration on it.
00:19:11.359 --> 00:19:35.119
And this is really important in psychiatry because psychiatric neuroscience is particularly vulnerable to the temptation to turn correlations into causations, to turn correlations into mechanisms of action of a disease, and to turn correlations into mechanisms of therapeutic promise.
00:19:35.519 --> 00:19:45.200
That's not appropriate, because biological plausibility is not the same thing as demonstrated clinical efficacy.
00:19:45.519 --> 00:19:57.200
So our job, and we will all get better with it as time goes by, and not that we are naysayers or we are curmudgeons or always looking for the negative.
00:19:57.359 --> 00:19:59.200
No, that's not who we are.
00:19:59.599 --> 00:20:18.000
But the art of distinguishing evidence from the interpretation of the authors is very important if we want to become a particularly high-quality scientific paper reader in the world of psychiatry.
00:20:19.119 --> 00:20:23.359
Let's move on to habit number 10, and this is one of my favorites.
00:20:23.680 --> 00:20:32.960
That is, read the limitations section of the paper before you read the discussion section of the paper.
00:20:33.279 --> 00:20:41.039
Try that out next time you read a paper and just watch what a dramatic impact that has on our understanding of the paper.
00:20:41.279 --> 00:20:45.680
This is a small habit, but it can have a very large impact.
00:20:46.400 --> 00:20:51.119
Because before reading the author's interpretation, read the limitations.
00:20:51.440 --> 00:20:56.160
Then ask yourself, what limitations have I identified on my own?
00:20:56.400 -->
00:20:58.160
What biases remain?
00:20:58.400 --> 00:21:01.759
What alternative explanations deserve consideration?
00:21:02.079 --> 00:21:12.799
And perhaps the most question, most important question of all, what would a skeptical but fair-minded clinician conclude from these findings?
00:21:13.359 --> 00:21:17.440
And only then read the discussion section.
00:21:17.920 --> 00:21:25.680
I will be honest with you, discussion sections very often, I think, a slight exaggeration of the findings.
00:21:25.839 --> 00:21:35.039
As I said before, authors have kind of a, I think, unconscious, subconscious desire to put the most positive spin.
00:21:35.279 --> 00:21:36.880
That's just human nature.
00:21:37.119 --> 00:21:41.119
Most of us do that when we write papers, me included.
00:21:41.359 --> 00:21:52.160
So if you ever read a paper of mine, you should not automatically assume my discussion is in fact the correct one.
00:21:52.400 --> 00:21:59.759
What you ought to do is read the limitations first, read the methods first, look at the data.
00:22:00.480 --> 00:22:03.920
First, and then you turn to the discussion section.
00:22:04.079 --> 00:22:14.079
Not because you're a cynic, but because you are a disciplined curiosity seeker in terms of psychiatric literature reading.
00:22:14.480 --> 00:22:14.799
Okay?
00:22:15.680 --> 00:22:16.240
Very good.
00:22:16.400 --> 00:22:21.759
Thanks for letting me share this with you from the very bottom of my heart.
00:22:22.000 --> 00:22:32.799
We move on to habit number 11, which is use artificial intelligence as a reading assistant, not as a scientific authority.
00:22:33.440 --> 00:22:36.000
Folks, I came to AI pretty late.
00:22:36.319 --> 00:22:40.160
I almost got shamed into it by my younger colleagues.
00:22:40.319 --> 00:22:49.680
Some of my colleagues who are 25 plus years younger than me told me I really need to get on the bandwagon, and I have now.
00:22:50.000 --> 00:22:55.359
And I find AI to be useful, but I really want to offer you this caution.
00:22:55.680 --> 00:23:00.799
Think of it as a reading assistant and not a scientific authority.
00:23:01.200 --> 00:23:13.119
So AI can help us summarize complex papers, it can clarify statistical concepts, it can help us compare study design, it can help us identify questions worth investigating.
00:23:13.279 --> 00:23:21.759
But there is a really important danger here because a beautifully written summary can create an illusion of understanding.
00:23:22.240 --> 00:23:24.160
Careful, careful.
00:23:24.480 --> 00:23:29.440
AI can't do critical understanding of data.
00:23:29.759 --> 00:23:31.759
My recommendation, therefore, is simple.
00:23:31.920 --> 00:23:33.839
By all means use AI.
00:23:34.480 --> 00:23:38.720
But by all means, by all means, do your own reading.
00:23:38.880 --> 00:23:54.000
AI will never be able to match up with you and your abilities in terms of critical questions that need to be asked, but as importantly, how to draw clinical conclusions from it.
00:23:54.240 --> 00:23:57.920
So AI should amplify our scientific judgment.
00:23:58.160 --> 00:24:00.160
It should not replace it.
00:24:01.119 --> 00:24:10.559
And then finally, habit number 12, which is always end with the most important question.
00:24:10.799 --> 00:24:22.799
So after examining the methods and results and limitations and interpretation and discussion section of the paper, I would encourage you to ask one final question.
00:24:23.759 --> 00:24:32.000
What did I get from this, if anything, that'll help me do things differently on Monday morning?
00:24:32.400 --> 00:24:37.519
And that, of course, is a metaphor for what can I do with this tomorrow?
00:24:39.839 --> 00:24:49.920
So the question asking you to perhaps pose to yourself, which is this question connects scientific literacy with clinical wisdom.
00:24:50.160 --> 00:24:55.680
So the study perhaps provides compelling evidence, but is it practice changing?
00:24:55.920 --> 00:24:57.519
Is it something to keep in mind?
00:24:57.680 --> 00:25:00.000
Is it something that I need to watch out for?
00:25:00.480 --> 00:25:03.519
Does it challenge assumptions I've had for years?
00:25:03.839 --> 00:25:05.279
Should I change my mind?
00:25:05.519 --> 00:25:07.440
Should I not change my mind?
00:25:08.000 --> 00:25:09.759
All of these are important.
00:25:09.920 --> 00:25:13.599
We don't read scientific papers for the pleasure of reading.
00:25:13.759 --> 00:25:15.599
We do that with fiction.
00:25:15.920 --> 00:25:17.359
This is not fiction.
00:25:17.680 --> 00:25:18.960
This is different.
00:25:19.359 --> 00:25:32.079
Every single paper after I have performed due diligence, the question I need to ask myself is what is the clinically relevant learning I got from it?
00:25:32.319 --> 00:25:33.839
Is this interesting?
00:25:34.160 --> 00:25:35.599
Is it promising?
00:25:35.920 --> 00:25:38.079
Is it not yet practice changing?
00:25:38.319 --> 00:25:40.160
Is it practice changing?
00:25:41.119 --> 00:25:43.839
That is scientific maturity.
00:25:44.079 --> 00:25:47.440
That's how you and I read a scientific paper.
00:25:47.759 --> 00:25:51.519
So I'm glad I got to talk to you about those 12 steps.
00:25:51.920 --> 00:25:53.440
And maybe I'll do this.
00:25:53.599 --> 00:25:57.119
I'll give you my five-minute scientific reading checklist.
00:25:57.359 --> 00:26:00.160
So here's a practical framework you could use.
00:26:00.400 --> 00:26:06.640
So before you accept the conclusions of a scientific research paper, perhaps ask yourself, question one.
00:26:07.039 --> 00:26:09.680
What is the question being asked by this paper?
00:26:09.920 --> 00:26:14.240
What exactly were the investigators trying to ask and establish?
00:26:14.640 --> 00:26:16.559
Number two, the design.
00:26:16.960 --> 00:26:20.319
Was the study designed well enough to answer that question?
00:26:21.759 --> 00:26:24.319
Third question, the population.
00:26:24.720 --> 00:26:27.119
Do these patients
resemble my patients?
00:26:28.240 --> 00:26:30.480
Question four, the results.
00:26:30.720 --> 00:26:35.519
What is the magnitude and precision of the benefit demonstrated?
00:26:35.759 --> 00:26:36.960
And the same thing.
00:26:37.119 --> 00:26:43.599
What is the magnitude and precision of the harms generated for patients in the study?
00:26:44.480 --> 00:26:48.000
Question five, what is the meaning of this?
00:26:48.319 --> 00:26:52.559
Are the conclusions justified by the actual data?
00:26:53.440 --> 00:26:57.359
And number six, of course, the most important one, application.
00:26:57.839 --> 00:27:02.880
Should these findings change what I do clinically?
00:27:03.359 --> 00:27:16.400
Six questions you can ask yourself in five minutes, perhaps ten minutes, we can absolutely do that as we read scientific papers because that is the lifeblood of progress in our field.
00:27:16.720 --> 00:27:19.440
So a final thought in the last minute or two.
00:27:19.920 --> 00:27:25.279
Look, folks, scientific reading is an act of intellectual humility.
00:27:25.599 --> 00:27:39.680
And the longer I practice psychiatry, now that I am well into my 40th year in psychiatry, the more convinced I am that scientific wisdom is not defined by how many studies we have read.
00:27:39.920 --> 00:27:44.720
It is instead defined by how carefully we interpreted them.
00:27:45.039 --> 00:27:47.119
It does require curiosity.
00:27:47.359 --> 00:27:51.920
It does require some measure of statistical literacy, but not a whole lot.
00:27:52.319 --> 00:27:57.119
But it does require the courage to question prevailing assumptions.
00:27:57.440 --> 00:28:04.640
And perhaps most importantly, it requires the humility to change our minds when better evidence emerges.
00:28:05.279 --> 00:28:15.759
So my advice in this generation of information overload, increasingly sophisticated research, AI-generated summaries.
00:28:16.000 --> 00:28:17.599
Here's my advice.
00:28:18.400 --> 00:28:20.000
Read widely.
00:28:20.640 --> 00:28:22.880
Question carefully.
00:28:23.759 --> 00:28:26.000
Think independently.
00:28:26.640 --> 00:28:32.799
And be ready to change your mind only when the evidence warrants it.
00:28:33.200 --> 00:28:38.720
And let's remember our patients don't need to be impressed by the latest research.
00:28:38.880 --> 00:28:42.240
They need to understand what it actually means.
00:28:43.279 --> 00:28:47.920
So thanks folks for letting me start this important conversation with you.
00:28:48.160 --> 00:28:52.160
I would love to hear from you in terms of how you read the paper.
00:28:52.480 --> 00:28:59.039
How do you bring meaning to the scientific facts that emerge at us?
00:28:59.599 --> 00:29:06.319
And I wish you the very best as you continue developing your positive psychiatry practice.
00:29:06.480 --> 00:29:16.880
And of course, and of course, reading high-quality papers, understanding them well is a foundational aspect of becoming a better clinician.
00:29:17.119 --> 00:29:17.519
Right?
00:29:17.680 --> 00:29:19.119
Don't you agree with me?
00:29:19.759 --> 00:29:23.039
With that, folks, I wish you the very best.
00:29:23.279 --> 00:29:30.000
This is Rakesh Jan, and it's also a pleasure to have you on this podcast episode of Texas.
00:29:31.359 --> 00:29:31.759
Goodbye,