On Reporting Quality Reproducibility - Tips for Success with Progress in Orthodontics & Other Journals

00:00  Slides 1-2

Again, I have no conflict of interest, except from the fact that maybe that I'm the editor of the journal.

00:12  Slide 3

So, Progress in Orthodontics is an open access journal, which means that all your papers will be available to anyone who visits the site.

You don't have to pay, you don't have to buy them, you don't have to have a subscription.

00:28  Slide 4

Since a couple of years ago, the journal used to be free to publish.

Now you have to pay, and it's quite expensive, I think.

But there is funding available, discounts and waivers.

And you have to go to the site, because it depends on the country and so forth.

Maybe some universities have a contract with the publisher, which is Springer, and then they can get a discount.

And also, we have four papers per year, waivers, we can do them for free.

Those are usually invited publications, but you can also submit and request that.

But it's important to do it in the beginning, because it makes the process much easier.

And once you get accepted, then you may be granted a waiver.

01:23  Slide 5

If you go into the submit a manuscript section, we have three types of publications.

You have a review, which is a systematic review.

We don't publish actually narrative reviews, so it has to be a systematic review.

We have a research article, which could be like a randomized clinical study, an observational study.

We also have studies, laboratory studies, but we don't focus so much on those and also you have brief reports.

01:56  Slide 6

One thing that I wanted to do, which I've done in the American Journal of Orthodontics since I started in 2011, was to introduce and enforce the reporting guidelines.

02:10  Slides 7-8

And why is that important?

Because you design your study, you execute, you publish it.

That means you had your patience and you used a lot of funding, maybe a lot of time, to create something which hopefully would be and should be valid and also useful.

Now if your publication does not have the information which is important to make it useful, then this is wasteful information.

And because you have to remember we focus on a systematic review, but a systematic review is populated by individual studies.

Therefore, if you don't have individual studies of high quality, you're not going to have a systematic review which you can have a lot of studies and solid evidence for which you can feel confident about.

So it's very important that you publish what you did correctly and transparently so people can evaluate your work.

Because then you can use this study for a systematic review, or you can get ideas for a new study.

03:25  Slide 9

And also you can draft healthcare guidelines.

And this is what I said before, reporting guidelines are not about methodology, how to conduct your study, or how they're associated with it.

They're about telling exactly the reader how you did and what you did.

This will allow me or anybody who's interested to make good assessment of the methodology and the risk of bias.

Which in turn will mean you will do correct interpretation of the study findings and also reduce waste and have optimal use of research.

Because bad research is not good, so we have to improve it.

04:12  Slide 10

This is a site, maybe you want to take a photograph of it, if you don't know it, it's called the Equator Network, and it has a plethora, has many guidelines for reporting guidelines.

04:27  Slide 11

And we focus mainly on those three, CONSORT for randomized clinical studies, STROBE for observational studies and PRISMA for systematic reviews.

04:40  Slide 12

If you go now into progress and you want to submit, let's say, an observational study or randomized clinical study, unfortunately the publisher doesn't allow me to make a lot of changes because I want to keep Springer journals to be all the same across all the disciplines.

So you can only find that, there's a hyperlink where you see it, okay?

It will take you to a document, where it says how to format the paper, because we use subheadings, it's not like introduction, methods, results and discussion.

We have several subheadings, for example, in introduction you see specific objectives and hypotheses.

In the methods, study design, participants, exposure outcomes and so forth, you actually need to go and add the subheadings in your paper and fill in the information. If you send the paper unformatted like that, it will come back to you and ask you to do that.

And it's not only important to put the information in, to fill in the blanks, but also to correctly do the information, because we also check that.

And the same for randomized clinical studies. And here I've written some more explanations on how,

For example, you know, to draft the tables.

And you know, when I show the effect of intervention, the confidence intervals, not just the p-values, but you have to consult with those links and see the information.

06:38  Slide 13

Perhaps you would like to take a photograph of that. That's a publication we did some years ago and we used an example, an orthodontic example, and we formatted this paper exactly how we think it should be. But it's all based on the CONSORT guidelines, but includes the subheadings that I told you.

07:00  Slide 14

And there's also this, this is from the American Journal of Orthodontics, because I wasn't able to do that in progress, that's why I'm showing that to you here.

I have an annotated example, which is a paper.

Okay, you see I have those keys in different sections, and if you go like that, and click on it, then it will explain to you why you need to include this information.

Okay, but that's on the American Journal of Orthodontics website, the annotated example.

So in every section you'll see the format of the paper, you'll also see the explanation.

And I think because that's an orthodontic example,

I think it's more helpful than going.

You can always go to the concert, but that uses mainly medical examples. and that's an adaptation from the consort into the orthodontic world.

08:01  Slide 15

And we also have this link for systematic reviews, and I'm not going to go into that, you can check it.

08:11  Slide 16

Another important thing is you have to declare your conflict of interest.

We request that also from the associate editors of the journal, if they want to publish something, they will have to declare if they have any, but the associate editors in the journal, and also if they have any affiliation with a company or if they're getting, the specific requirements that you can see on the website, you know, if you lecture for the company, if you get paid for a company or if the product you're introducing to the publication is funded by a company and so forth.

I'll declare that.

08:50  Slides 17-18

Another thing which I have not fully introduced yet is the data sharing.

Because when you do a systematic review, has any of you done a systematic review?

So one issue that you encounter is how the missing information, the missing data.

Therefore, if you publish your paper and you share your data, then anyone who wants to do a systematic review can take this information, the data, and use it for his meta-analysis.

And that's not always easy.

You think it's easier than it's easy, but it's not.

And that's all connected with good quality reporting.

Given the information that is associated with the paper, so somebody else can use this information to take it to the next level. That's the point.

09:50  Slide 19

Now we'll go a little more into and explain the main components of the CONSORT and what they include and why you should include them in your publication. So it'll be a little more detailed.

10:05  Slide 20

This is the update of the CONSORT, the standard CONSORT for a parallel clinical study. It was updated in 2025 and it was published at BMJ and other journals.

10:17  Slide 21

And we also published in 2017, an extension to the CONSORT 2010, which was the previous one, which is associated within person randomized designs, which is relevant to dentistry because we include split-mouth designs.

10:36  Slide 22

So the CONSORT consists of different sections.

As you can see, it has 30 main items plus some sub-items.

10:49  Slide 23

And you start with a title.

So in the title, you want to show that it's a randomized study.

Because when I do a systematic review, I want to be able to find quickly, if I'm looking for randomized studies, the studies that will be eligible for my systematic review.

So that's why you need to include that.

11:09  Slide 24

I'll skip the summary.

And then it's important maybe in the title to follow this PICOT, to include information about the participant, interventions, comparatives and outcomes in design.

11:19  Slide 25

And that's an example, for example, of a title, survival of bonded, survival, that's the outcome. with chemical, that's intervention A, or photopolymerization intervention B, over two year period of time, that's also important, a single center randomized clinical trial.

So that's an example of a title that includes all the PICOT components.

11:57  Slides 26-27

For the within-person designs, we would also require the extension, the inclusion in the title of the within-person design.

And it's important again, because if we're looking for split-mouth designs, you know, to include that, or if we want to exclude those, we can exclude them, because we know from the title, it makes it easy to filter through the publications.

So that's an example of that.

12:27  Slide 28

That's a new component now has been added, it's called open science.

It was in the previous CONSORT.

12:35  Slide 29

And it talks about trial registration and protocol.

Why is that important?

Because what happens often, you have a publication which was designed to examine a particular outcome.

But maybe the authors did not find something statistically significant, so they changed the outcome.

So you had a study which was designed for something, and then it ended up being for something else.

If you have a protocol, then you can always make a comparison between the protocol and the final publication to see if there are any important discrepancies.

Small differences could be found and can be explained, but bigger difference should be clarified or maybe they should make it possible to publish a paper based on this previous protocol.

And actually some of the high-end medical journals, they ask you, they require you to have a protocol along with your publication. They will not publish something unless you have a protocol.

13:37  Slides 30-31

If you go in the introduction, you want to give the scientific rationale and the specific objectives.

13:45  Slide 32

So your introduction will include a description of the problem.

You have to discuss the existing evidence and give the rationale for doing a study because it's only ethically acceptable and justifiable to do a new study.

if you have what we call equipause, which means a scientific uncertainty, uncertainty in the scientific community about the effectiveness of the intervention.

If there is something that we know the answer for, then you cannot justify ethically to do a new study. So you need to include that in your introduction, and you need to find, if possible, a systematic review, identify knowledge gaps, and then based on the knowledge gaps you have, you build in on the next study.

14:33  Slide 33

You do the cycle of knowledge. You start with a question, you do a systematic review, you identify what is missing, and then you do a new study and you update your systematic review.

14:49  Slide 34

And as somebody said, a single study is an island in search for a continent, which is the systematic review. So that's the cycle of evidence. Do a systematic review, identify what's missing, do a new study, update the existing systematic review.

15:10  Slides 35-37

If you go in the methods, there are also several items here. So one has to do about patient and public involvement, that's also a new item which wasn't available in 2010.

And this is important because you see in orthodontics a lot of big tables of cephalometric measurements, which are selected in a random arbitrary way, which may mean nothing to the patient.

So getting what is important outcomes for patients is an important step, and we need to focus on that.

And another idea is you need to standardize the outcomes, because if we have 10 different studies on a particular intervention, but we measure different things, then we cannot combine them in a systematic review.

And that's what we call the core outcomes and standardizing important patient, important outcomes.

16:18  Slide 38

In terms of the design, we want to know whether this is a parallel design, a split-mouth design, and this information is guiding you automatically as to what analysis you should do, because if you have a split-mouth design, you have paired data, so you need to analyze your data in a different way.

If you have a cluster design, again, you need to do different types of analysis, and I think I have just a little diagram.

16:46  Slide 39

So that's a parallel design where you have, let's say, 100 patients and 50 patients are randomized to take treatment A and another 50 patients are randomized to take treatment B.

Or you have a split-mouth design or a within-person design.

So we're missing two lateral incisors, and we're going to test, let's say, two types of Maryland bridges.

That's a different design.

you have to be knowledgeable about how to analyze because this will be different than the design above.

Or you could have what we call a cluster design and I'll explain in a bigger, in a magnified version.

17:25  Slide 40

You see you have white spot lesions on several teeth on both sides.

So if you make multiple measurements within the same patient, because here you have multiple teeth, you have what we call clustered effect.

And this has an effect on how many patients you need, because when you have multiple observations within a patient, you lose information. And also has to do with how you analyze your data.

If you analyze your data at the tooth level without accounting for the fact that those teeth belong to the same patient, you will get a high sample size, which will make your p-value very small, you'll find a statistically significant result which may not be genuine.

So this design could be actually a combination of a within-person cluster design. You could say I'm going to use one method on the right side for the white spotless, but I have multiple teeth, that's why I have clustering effects, and on the other side I will use another method.

And again, I have clustered effects, but because I'm using the same patient, it's a split-mouth design.

So this design requires a different type of sample size calculation and different analysis.

Perhaps you'll probably need to use like a mixed model.

You cannot use like a simple linear regression or a logistic regression.

19:00  Slide 41

The next item is eligibility of participants.

19:09  Slides 42-43

And we have an extension because in the split-mouth design, like an example I showed you before, if you're missing two lateral incisors, then you have a patient who has two sides which are eligible. But if you're missing a lateral and you're missing a premolar, then you cannot use this patient. The same you have to do endodontic treatment and so forth.

19:37  Slide 44

Or like that, if you have recessions, this could be a good example, you can do a split mouth design.

Perhaps you can use one intervention on one side and one on the other.

Maybe a pair of donors can tell us if it's a good example of it, I guess, because they're too close.

19:56  Slide 45

Now in terms of the interventions, you need to give exact details, how the interventions were applied between the two groups.

And the level of detail is as follows: you have to allow someone to be able to duplicate what you did, based on what you read in your paper.

That's how clear it should be.

20:25  Slide 46

And also you know how many times you apply the intervention and so forth.

The next one is what we call the outcomes.

What are you going to measure?

So we have usually one or two primary outcomes and maybe we have secondary outcomes.

So again, you need to give the details on how you made the measurement, which were the outcomes, how frequently you measure those outcomes and so forth, all the details.

20:56  Slide 47

An important component of any clinical study is to check effectiveness, but also to check side effects, because if something is effective but it's hard for the patient, you have to balance the plus and the minuses before you make a decision.

21:14  Slide 48

You need to show exact details of how you calculate your sample size.

You can't say I used the paper so-and-so that was published in 2010.

You have to give the exact numbers and the assumptions that you made, so that I can replicate the sample size calculation to make sure the number you came up agrees with standard formulas.

And why is it important to have a correct sample size calculation?

If you have too few patients, then your study is going to be futile, as we say.

It's going to be a waste because you're not going to show anything.

It will be an inconclusive study.

On the other hand, if you have too many patients, maybe you're exposing some patients to a therapy which is harmful.

So that's why you have to balance and get, based on logical assumptions, the correct sample size.

22:15  Slide 49

An extension for the within-patient design is that the advantage of the split-mouth design, when it's applicable, because sometimes it's not applicable, is that because you're using the same patient, you have decreased barriers.

This means that you will have, you are required to have fewer patients.

So in terms of efficiency, it's a more efficient design.

However, you have to be careful not to have what we call a carry-off across effect, which means, for example, if you're testing two mouthwashers, mouthwash A on one side, mouthwash B on the other side, if the effect of, if the mouthwash goes from one side to the other side, then this study is not eligible, it's not appropriate for a split mouth design. So you need to be careful with that.

23:08  Slide 50

So this is a, it shows, you know, the advantage of that, as I said.

23:14  Slide 51

The next step, the next item is the important item, all are important items, I guess, but this is what, you know, justifies the randomized clinical studies, how you did the randomization, the different components of how to randomize, there are methods that are acceptable, and methods that are not acceptable in terms of how you randomize your patients.

And I'm not going to go into details on that, but you need to explain accurately how you randomize your patients.

23:50  Slide 52

And then for the split-mouth design, if you have a parallel design, you have to randomize individual patients to A or B. But if you have a split-mouth design, you would have to randomize body parts.

Remember the example I showed you with the two missing teeth?

How should I decide which intervention will go on the right and which one will go on the left?

So you have to explain that also.

24:16  Slide 53

And then you have another level actually.

So this is with the eyes, which eye will get A, which eye will get B.

But you could have a second level of that and say,

You can randomize the eye drops first.

So I'm randomly selecting one of the two bottles.

So that's one level of randomization.

And the second level is I'm gonna randomly apply the drops that I selected randomly at the previous stage to which eye.

So two levels.

24:54  Slide 54

Then we have what we call a method, which is a part which is called allocation concealment.

This is not blinding, it's a method that we use to make sure that the patient or the investigator cannot predict what the patient would get.

And obviously the best way to do it, perhaps you've seen the opaque sealed envelope concept, but this is kind of easy to subvert.

You see this photograph of the doctor looking up.

Because what happens if I'm an investigator and I believe treatment A is better than B, if a patient comes in and I believe he will be a good candidate for B or for A, because subconsciously I want to prove my point, I will subvert the randomization and put him where I think he will be best so I can prove my point.

So the best way to do allocation consumer will be to use like a computer system, like an off-site place where you actually log in and you get the randomization automatically without you having to interfere.

26:07  Slide 55

Then we use blinding and how was blinding applied, if it was applied and how it was achieved. So that's another item.

26:17  Slide 56

Then we can go into the statistical methods, and we have to explain how we analyze our data.

And obviously, depending if we have a cluster design, if we have a split-mouth design, we have to use appropriate methodology.

And those are mistakes that we often see.

We've done several publications of dental and orthodontic papers, and we found that those methods are not used appropriately.

And that's why it's important to understand them.

26:49  Slide 57

Another issue is what happens with missing information.

If you have, let's say, a longitudinal study, and you randomize 200 patients, and then you finish your study and then you lost like 80 patients, now you have 120.

It's important to show what happened to those patients and why those patients were missing.

Is it because they moved away or is it because of the intervention?

Because you randomize your patients to make them similar in both known and unknown parameters in the beginning.

But if you lose a lot of them, and if you lose them, if the patients you lose in Group A and Group B are different in baseline characteristics, then you messed up your randomization.

You have what we call post-randomization bias.

Your treatment groups are not the same anymore because you lost patients with different characteristics.

And that's why it's important to know that.

27:48  Slides 58-60

Now we'll move to the results.

And you need to include a participant flow diagram.

27:58  Slide 61

I'm sure you've seen that.

And that's an easy way to visualize the flow of your patients

and whether you have missed any patients and why you missed them.

If somebody moved away, if you lost a few patients because they moved away, they changed job or something like that.

That's not an issue of bias.

But if you lose them for other reasons, and now your patient groups are not the same anymore, then you have bias, and then the conclusions and the results will not be valid.

28:31  Slide 62

For the split-mouth design, we have a particularity because you have the same patients.

So you have the two groups within the patients.

So you see in this one, you're randomized 71, so one eye gets treatment A, the other eye gets treatment B.

While on this one you have different patients.

So that's how you have to format your split-mouth flow diagram.

29:04  Slide 63

You need to show a baseline table.

29:07  Slide 64

That's an example of a baseline table. It shows you the number of patients or teeth, whatever that is.

And then it shows you the baseline characteristics of those patients.

You don't want to do a statistical test.

That's a common mistake.

You randomize and then you do on the baseline table statistical test to see if they're similar.

That's a problem for two reasons.

One is because if you have a lot of variables, you get a false positive p-value.

And the other one is that because you expect some variation in baseline characteristics, and that's normal.

And then another reason would be also if you have a small number of patients and you have big differences between the patients, baseline characteristics like age and gender and so forth, you may get a non-significant p-value just because your sample size is small.

And this will give you the fast impression that your patient groups are the same, when in reality they have big differences.

So that's one group and that's the other group.

30:18  Slide 65

Now you have a particularity for the split-mouth design, because some of the characteristics are the same, because you have the same patient, like age, gender, and so forth, are the same.

But some other ones are different, because they have to do with the body size, like the eyes or the teeth and so forth.

30:39  Slide 66

And you have to give the information, how many patients you had, how many patients you analyzed, remember the missing data and so forth.

30:54  Slide 67

And you have to show this table, you have to show effect estimates, we don't accept only p-values.

Because the effect estimate shows you in terms you can understand, and you are a clinician, so you understand if I tell you one intervention, decrease the over-jet by 5 mm, the other one decrease by 10 mm, you understand this difference of 5 mm is clinically important.

And if you have a small number of patients, your p-value will not be significant.

But you want to see actually the effect of the intervention, that's the effect of 0.3 that you see, and the confidence interval, the uncertainty around that effect.

So that's an important item to include in your paper.

31:44  Slide 68

Now, if you have binary outcomes and what I mean a binary outcome, if it's like a yes or a no, I lost a bracket, I didn't lose a bracket, it's important to show absolute and relative effect.

And I'll explain to you what this means.

32:03  Slide 69

So let's say, if I'm taking intervention, the risk of dying is 1 in 1,000.

And if I take the placebo, my risk of dying would be 2 in 1,000.

If I use a relative effect, that says the risk of dying in B is twice as high compared to 1.

So you can actually fool someone with a number, with a statistic, because you tell them, you know, if you don't take this and take the placebo, you have twice as many chances of dying.

So it's important to take the intervention.

But if you actually see the absolute difference is only 1 in 1,000, which clinically doesn't mean anything.

So that's why it's important to understand and show what is the baseline that you're starting from.

Also it's different, if you go from 1 to 2, it's twice.

But also if you go from 40 to 80, that's also twice.

But one is a difference of 1, the other is a difference of 40.

And you miss that if you use only the relative effect, the risk ratio.

33:29  Slide 70

Again, similar things that we want also for the split-mouth design, but we also want to use the correlation coefficient.

33:41  Slide 71

As I said before, the reason we want to use the correlation coefficient is because we can use that value to make our new design, the impartial design, more effective because the correlation coefficient will reduce the number of patients we need.

And if we know that, then we can make this adjustment to include fewer patients and have a study with a similar power or adequate power.

34:09  Slide 72

You have to present also the harms and the side effects that you have.

34:13  Slide 73

Now, in terms of the discussion, you have to balance the benefits and harms and consider also the relevant literature.

So you give your conclusions, you say, discuss the positives and negatives, and also you talk about how those conclusions fit into the continent.

Remember the island looking for a continent, how these findings that you have are related to the existing and scientific evidence.

34:41  Slide 74

And of course you have to also discuss what limitations and possible biases you have. that's a section which is often overlooked, because if you discuss any biases, you think that the editor will reject your journal, but it's good practice, a good reporting practice to discuss the possibilities of bias.

35:05  Slide 75

Common mistakes that we have encountered in the reporting,

And I think also, you've done some studies on split-mouth designs and limitations of reporting.

One is what we call no publication.

Not all studies are published.

So if 100 studies are conducted in a topic, maybe 50 of them are published.

The negative effect of that is the following.

When you do a systematic review, you can only find the 50 out of 100.

So you're already biased because you don't have all the information.

In addition to that, the studies that are usually published, unfortunately, are the ones that have a statistically significant result.

So you're also biased on the positive.

So the negative studies are not published, you see only positive studies, so you do a systematic review and you're biased on the positive.

You have incomplete reporting of trial elements.

So I gave you the list of those 30 items and the sub-items, and I showed you what needs to be included, I think you can understand this point, that the information is missing.

You have also what we call misleading reporting or SPIN.

SPIN refers to the following:

I have non-statistically significant results in my study.

I don't give the estimate or the confidence intervals,

I say I have non-statistically significant results, and then I go on to conclude,

but there was a positive trend for the intervention A compared to intervention B.

And if I had more patients, then I would have been able to detect this trend and have a statistically significant result.

You see that in publications, you also see that very often in the media, and that's an incorrect conclusion because if you design your study again, it doesn't mean with the same intervention we get the same result.

Because what we do, we do inference. We take a sample from the population and hopefully, if it's well designed and unbiased, we make a conclusion which we can infer for the entire population.

But if you do many studies, you have a random error because it fluctuates, because you take different samples. So you can never be sure if what you find is what is going on for the population.

The only thing we can do is we can increase the sample size and also we can apply methods to reduce the bias. But bias is not something you can measure, it's not like 10 millimeters, one meter, you can measure that, you just apply methods, empirically tested, that will help you to reduce bias, but you never know how much bias you have left.

Another thing is what we call selective reporting of the analyzed outcomes. Maybe we looked at different outcomes and we decide only to discuss or highlight the ones that have a statistically significant result. That's common and you can reduce that if you have a protocol, because then somebody can compare what you intended to do and what you ended up doing.

Incomplete reporting of elements of precision, that means we're missing the confidence intervals. That's a very common issue.

And then the final, that's not an exhaustive list, by the way. That's just some examples, okay?

And I highlight the last one because I see it very often.

You read the publication and you cannot understand what the authors are trying to say.

You start with the objective, you go in the methods and the analysis and the conclusions, and those things are not aligned.

You see the PICOT, and then somehow all this thing is lost.

Then they analyze all kinds of different things, and then they highlight something else often that is not actually in the objectives, which were given in the top.

So that's an important issue, it has to be coherent, from the beginning to the end.

Don't try to make it complicated, make it simple.

When you give the publication to somebody who is not familiar with the topic, they should be able to understand quickly what you did.

And if you follow those principles, this will make your life easier.

39:57  Slide 76

That’s a study, an older study we did, and you'll see sections which were inadequately reported consistently. Sample size calculation was not reported. Randomization, blinding, intention to treat, this is related to how many missing patients you had and so forth.

Confidence intervals, the focus is often on the p-value.

Multiple testing, a lot of testing is done.

You have an outcome, you don't find something statistically significant, then you'll say, okay, if it's not significant, maybe my paper is not interesting.

So then you start to say, okay, I'm going to see whether it's more effective in female patients or male patients.

So then you do subgroup analysis, or maybe you do age groups, or maybe you say, I'm going to test in high-angle cases versus low-angle cases. So you start a number of multiple tests which are not pre-specified, which will give you possibly false positive results, and will make you susceptible to selective reporting, because you don't have something statistical, so you start digging your data until you find something.

And when you find something, you know, you say,

"Okay, I found it, statistically significant."

That's not the way to do things.

And those are the sections which are usually well done,

which is the hypothesis, objectives, data collection, intervention, and so forth.

41:38  Slide 77

It has been postulated, I don't know how accurate that is, okay, and it would depend also on the specialty, that from all the studies that are done, 50% of them are not published.

From the remaining 50%, half of them have sub-optimal reporting, not good reporting, which could mean that you cannot use them.

And 50% of the remaining ones have major design flaws.

So you start with 100 studies and maybe 12 of those are valuable.

That's the point.

42:21  Slide 78

So 85% is wasted, over 85%.

I cannot verify this number, but that's the point.

You see how it's multiplicative, the effect.

You get multiplied.

42:33  Slide 79

And that's a paper we published some time ago.

We talk about waste and the different components of it.

42:41  Slide 80

So a waste reduction and optimal use of research is our responsibility.

That's why we need to teach good methods, to understand the good methods and apply them.

It's something that we owe to our profession, to the people that make it happen, because to conduct a study you need funding, you need to get patience, you need to take time, you need to follow them up and so forth.

So you need to make a study which will be useful and valid.

And above all, because we do research to help our patients.

And if we're not doing good research, we're not helping our patients.

So we owe that to them also.

43:25  Slide 81

And I think that's the final picture.

So good reporting is not an optional extra.

It's an essential component of good research.

And this is, I don't know, you know this guy, huh?

He passed away some time ago.

And that's at the CONSORT group meeting in Paris.

He was quite influential on that.

He's a well-known statistician, his name is DG Altman.

He has a nice book.

And he was a big proponent of reporting guidelines.

And himself and some other people from other parts of the world, they've started the CONSORT guidelines.

44:08  Slide 82

And before I finish, we have a congress, it's a biennial, every two years, the last one was in Berlin, and the professors from the university came, they honored us with their presence.

So the next one is in Athens in November, and the topic is Artificial Intelligence in Oral Health.

And actually the station has opened.

The program is, and you can also submit an abstract because we have five or six keynote presentations and then we'll have in between either oral presentations or poster presentations.

It starts on a Thursday afternoon, it's all day Friday and then it finishes Saturday afternoon.

I know it's far away, but maybe you can spread the word.

Okay, thank you. So that's the end of this part.