Nynke Scherpbier - Choosing wisely and guidelines

Transcript

This transcript is AI generated and may contain errors.

It's so lovely to be here and to talk with you about choosing guidelines wisely.

And guidelines may sound boring, so I hope I do not over-treat you.

I can see that they are not boring if you think about guidelines.

So you've already heard who I am, but not why I'm here.

Maybe I think that the committee invited me because I was the vice chair of the guideline authorisation committee in the Netherlands for the Dutch GP guidelines, which are a bit famous I hope.

And I'm also a chair of the guideline revision on chronic kidney disease which is ongoing and I do that for the primary care part and there will also be a revision of hospital guideline and also vice chairing that one so that's a nice collaboration between primary and hospital care.

So this is the College of General Practitioners in the Netherlands, and I'm very proud that they're really supporting the development of guidelines in a wise way, but it's not easy.

I have no conflict of interest to declare, I heard yesterday from Margaret that we should mention that, But I do have something to disclose in light of sustainability.

This is the city I work in, Groningen.

You're always welcome to visit it, it's a lovely city.

It's in the northern part of the Netherlands, and when I was planning my trip to come here, I thought, okay, should come by train.

And I looked this up in a Rome to Rio app, which told me that it would take almost 14 hours to get here.

I though, well, no way, not going to do that.

So I planned a trip by plane to Stockholm.

First, of course, the train to Amsterdam airport and also a train from Stockholm to get here adds up another four hours.

And you won't believe it, but this is also almost 14 hours to come here.

But it was three minutes faster than the tram.

It really made me feel sad and then I didn't even tell you that there was a problem with the airplane and we had one hour delay.

So guidelines and sustainable tomorrow.

I think there can be a mismatch between them, but it may also help us for the sustainable tomorrow.

So a very short wrap up and three slides on the big why of this.

You'll know that costs for healthcare are rising.

I present here the costs in your Nordic countries and also the Netherlands, and it's everywhere, it is the same game.

It's rising and rising, and not only the cost, but also the workforce.

So in 2018, one in seven professionals were working in the healthcare sector.

And if we continue this way, in 2040, 1 in 4 professionals will have to work in healthcare and of course that is not sustainable.

We need teachers, maybe we need even more teachers than Doctors, I guess, and we need policemen, so it is impossible to carry on like we do.

And we have already heard from Stefan the healthcare climate crisis.

There is a climate-crisis influence in healthcare, but where we should also be worried about is that healthcare is influencing the climate crises.

So in short, there is too much medicine and I created this little monster with AI.

I'm sorry, that also takes energy and water.

And sometimes it will pop up to remind you that in my speech too-much-medicine is the topic.

So could guidelines help to prevent too many medicine?

They may.

But guidelines are not a cookbook and students and young GP trainees often think it is a book.

But we need to learn how to handle the guidelines wisely, which is not that easy.

You have to know a bit how they come to a stand and also how you can decide not to follow a guideline.

This is our profession, actually.

So as I already said, there sometimes is a mismatch between the guideline and the real world.

Why?

Guidelines are based on trials, and a person in front of you in the consultation room is never this male person 40 years old with only one disease.

You recognize that, I guess.

And the guidelines may promote defensive medicine, the fear of missing out a disease, of not following the guideline and having legal consequences of that.

Guidelines often say do and hardly ever say, do not.

So these are all reasons that the guidance will not always help us.

They may promote too much medicine.

Here's the little monster.

And I will also show you some angels.

I'll try to tell you a bit how you can produce a guideline wisely and most importantly to use it wisely.

And this will not be a cookbook, actually.

It's not that easy.

And I will also show that they may prevent too much medicine.

In the last weeks, I've been thinking a lot about this topic and I had a lots of thoughts and thought, let me just explain my thoughts to you.

I hope you remember something out of that and take it away with you when you use a guideline or maybe produce a guidelines.

I will have four topics.

How we define diagnosis, how we diagnose disease, CRP testing, How is screen for disease and also where we defined disease.

Very, very important.

And to start with how define diagnoses.

You should realize that definitions really, really matter and there is something going on we call diagnosis expansion or disease-mongering.

I like that word.

And it was already in 2011 that the British Medical Journal had this journal where they draw attention on the need of a new deal on disease definition.

Because we have pre-hypertension, we had irritable bowel syndrome, well we've lots of diseases that maybe are not diseases but concepts.

For example, ADHD.

There is a definition for ADHD in the DSM manual, but it has changed over time.

And here you can see the prevalence of ADHD related to the development of these new definitions.

You see, the prevalent is going up for the simple fact that we have changed the definitions, so definitions really matter.

And what we usually think, what you should realize is that there are symptoms which are troublesome and we put these symptoms together and then we make the diagnosis ADHD.

But it's not a fixed biological entity.

It's the other way around.

it is not that ADHD as entity is causing symptoms.

This is a very important thing that we do together.

And this is called reification.

Reification is a difficult word, but it means creating things, making a thing out of a collection of words.

So that's very good to keep in your mind.

As this ongoing, the number of ADHD cases in general practice has gone up.

And also the number of prescriptions.

On the left side you see 2018. And the more brown, the prescriptions of ADHD medication, you can see they do very well in Sweden.

A few years later, with the changing DSM criteria, it's getting even more.

Also the Netherlands get along with it very, very very.

Which is not only a bad thing because ADHD medication is related with a reduced risk of several outcomes like self-harm, traffic crashes and crime.

But Lee and colleagues have produced interesting papers on the effects of this medication.

And this is a bit of a difficult slide, but it shows that during the years, the positive effect of medication has gone down.

So look here down at the crime numbers in 2006 till 2010. The benefit of medications was on a good side, the benefit was getting less.

It was still benefit, same for traffic crashes, unintentional injury, while not really clear in self-harm.

This kind of shows that if you prescribe medication to a bigger group, which is maybe not as badly affected, As before, the effect of medication will get less.

So this is completely in the line of too much medicine.

You see the little monster.

That's my first sub-conclusion.

This broadening of definition, diagnosis, expansion, promotes too many medicines.

My second thought is about how we diagnose disease.

And you will know these numbers of the diagnostic procedures that are going up.

For example, CT scans in the Netherlands, they're rising tremendously.

And these are numbers from the Regional Medical Journal of the number of laboratory tests that go up tremendously, which also is not only a bad thing.

Early diagnosis and treatment can lead to a better prognosis.

But we all know that there will be a lot of false positives leading to diagnostic procedures, worrying people, also taking a load of manpower.

And the big, big question is whether early diagnosis and treatment really leads to a happier life.

But that is a too big topic for today.

It is good to realize that how we define what is normal.

Here you see the distribution of cholesterol in the population.

So it's a normal distribution like we call it.

And on the left side you'll see if the threshold for too high cholesterol is at this point.

And we decide that the threshold should be lower, like we are kind of pushed to by cardiologists, by medical firms.

And it goes a bit backward, even not a lot, but you see that many more people will fall in the group of too high cholesterol that may need treatment and with all the side effects.

So our definition of normal is at the base of creating too much medicine.

And I want to illustrate this with a very small exercise.

I didn't bring any prizes, so there's nothing to win, but just to try.

Say the population risk for cardiovascular disease at your age is 1%. And somebody is coming up with a new risk factor.

And if you have this risk vector, your increase risk for CVD will increase with 30%. So then the question is, if your risk is positive, what is your for risk CVT?

Could you please discuss that with your neighbor?

Okay.

As a teacher, I learned that I should wait until the rumor is going down.

It is not, but I would like to continue.

So if 30% increase is a relative risk and your absolute risk is 1%, the new risk will be 1.3%. Who had the right answer?

Wow, that's half of you.

That's good.

Price is for all of But that is not very impressive, isn't it?

30% sounds impressive.

But if you see what it really means for this individual, we should realize whether this is worth all the trouble going into it.

Understand me well, it's not that I'm promoting to stop being a doctor.

It's just an illustration.

We could call this risk factorology.

And we mix disease and risk factors more and more.

I think with AI, we can find so many risk vectors in the future.

We should be really wise to discriminate it from disease, and know how to handle that.

If we try to lower already low risk, there will be hardly any positive benefit.

So I come to my second sub-conclusion with this rise in testing.

Of course, there will be a rise of known risk factors, but also in false positives and in problems with low absolute risks, too much medicine.

I'll come on to the third topic about screening.

Maybe you've heard about this, that CT screening in heavy smokers leads to a lower death rate in people that have been screened.

A lower deaths rate based on lung cancer.

It is 76% lower than in the control group.

This was a randomized control trial.

So this paper has led to screening programs for heavy smokers in many, many countries like in the US, UK, Taiwan.

And in The Netherlands there are also many people that push that we should do this lung cancer screening in heavy-smokers.

But now there are some extra research programs ongoing because the question is if there is a lower chance to die from lung cancer, of course that is important.

But how is the mortality from other diseases which may also be smoking related like cardiovascular disease?

Wouldn't it be better to put your money in the prevention of smoking?

So this is a very, very fundamental discussion.

We should not run after these big screening programs without thinking well what we are doing.

So this is a very fast sub-conclusion, but these studies on effectiveness of screening programs, they should study all cause mortality and not only dive into the specific disease you are looking for.

And then very much a GP thing, maybe a bit more difficult, but maybe the most important part of my speech today is where we define disease.

The setting where you work really matters.

On the left side you see the GP setting of the academic practice in my department.

And on the right side, you'll see a hospital setting in the hospital where my departments is based in.

There is a big difference between the two.

I think patients often don't realize and doctors often do not realize.

And I want to illustrate this with a new test.

Imagine there is a test for inflammatory bowel disease, IBD.

It's a wonderful test, sensitivity is 90% and specificity is also 90%. And my colleague in the hospital, she is a gastroenterologist and she's really enthusiastic about this test and says, okay, I'm using it very often because now I can diagnose IBD without doing a colonoscopy and you should do that in general practice also.

Why is this not a good idea?

In her setting, I warn you, this extreme example, just to show what is going on.

In her setting, the prevalence of IBD is, say, 40%. She's seeing a lot of patients with IBd.

And if she is doing this test in 1,000 patients, 30% prevalence means 400 patients will have IBD in her setting and the rest will not have IBT, 600. And if the real positive is 90%, because the sensitivity is ninety percent, out of these 400, 360 will test positive.

And out of the 600, 60 will test positive because there will be 10% false positive.

This means that there is a total of 420 people testing positive, 360 of them have the disease.

You still follow me?

And that means that the positive predictive value of this test is 360. So the real positive divided by 420. That's all patients that tested positive.

The positive predicted value is 86% in her setting.

Wonderful test.

But now, in the GP setting, Prevalence of IBD 2%. I will do the same exercise with you.

I would not ask you to do it yourself.

1,000 patients, prevalence 2%, 20 will have the disease and the rest will not.

980. This test, 90% will test positive, and 90 of 20 is 18 patients will be tested positive.

10% will be false positive out of the 980 is 98. The total is 116 patients will test positive in your GP setting, which means 160 test positive, 18 of them have the disease, and the positive predictive value of this same very test is 18 divided by 116 is 15.5%. I feel like I'm a magician.

This is the same test, but it behaves differently in a different setting.

And this is one of the cores of our profession, actually.

And it's very important to realize this, and that is why we need GPs in our guideline committees.

We need teaching Gps and not flying in medical specialists all the time, although they're very useful, but they do not always understand this difference in prevalence, which is called the prevalence fallacy.

So my fourth conclusion is that clinical reasoning is key in our profession.

Our students need to learn this and feel this, what it means.

And you have to know things like, What is the pre-test probability that a person will have the disease?

And what is added value of a test?

If there is no added values, don't use it.

So those were four of my thoughts I had.

And it's not that I can one and one translate it into guideline development and guideline use, but I'll make a small try to wrap it all up.

What does this mean for guideline and development?

The most important thing, in my view, is who is in the guideline committee.

We have the tendency to invite super, super specialists on the guideline committee.

But we need generalists, we needs GPs, We need economists, need patients on a guideline Committee.

And we do this in the Netherlands.

I do know that you do that in some countries, but that very often you have to same guideline for a hospital and for the GP setting, which I think I prove now is not a very good idea.

You need separate guidelines and they should be based on studies in primary care.

So we also need funding for studies and primary.

This is so important.

And I will explain you a bit about the GRADE methodology, which is a long word, grading of recommendation, assessment, development, and evaluation.

This is an international methodology that helps us to get from the evidence to the recommendations.

That is important step, but it's also a nice step to look at when you want to dive a deeper into a guideline.

And based on this paper in the BMJ on the great methodology, I want to illustrate some parts of the steps, which is too much to tell everything about it, but I will just point out some very important steps.

I've already talked about how important it is that the guideline committee is brought.

And a very, very important step before you do anything when you are in the guideline committee that there is consensus on what are the relevant outcomes.

It's not that you dive into the literature and the evidence and then make a synthesis and say, okay, this is the guidelines.

No.

Before you...

seek in the literature, you define what are relevant outcomes.

So to illustrate this with the chronic kidney disease guideline committee I'm now in, we had big, big discussions because yeah, what is a relevant outcome?

Of course, end-stage kidney diseases is relevant.

There is no debate about that.

And also cardiovascular disease is also relevant but is also relevant, like the nephrologists say, a decline in kidney disease in the renal function over time.

Is that a relevant outcome?

And, well, they convinced us that it is relevant but only a rapid decline because we all have this slow decline, of course, in kidneys function.

So this identification of outcomes is key and also to rate what you think is important and should have a heavy weight.

And then finally after reading the literature you come to an important step from the evidence to the recommendation where you also weigh equity, costs, feasibility and, nowadays, resources.

What does it mean for the planet, this recommendation?

What it does mean to the number of professionals we need for this guideline recommendation.

So I really want to point out this step and it's interesting to look it up in one of your guidelines when you doubt about the advice.

How did they come from the evidence to the recommendation?

And I think in the future this step will get more and more important.

And then of course you end up with recommendations and they can be here for weak or strong, but also against.

So the do not that are told in the beginning of my keynote is an important thing to also mention in recommendations.

And then for guideline use, which we all do.

So keep in mind that you have to discriminate disease, symptoms and risk factors.

You have weigh them differently.

And of course we should be excellent in clinical reasoning.

And as I already said, it's nice to check this evidence to recommendation steps.

And this is thuisarts.nl, so this GP online.

That's also an important step to make a translation from the guideline to something that is readable for patients with advice online So to conclude, I think guidelines and our profession, there is a kind of tension between them because the guidelines bring good things, they bring transparency, safety, efficiency.

But maybe the most important is, and we've heard that already many times at this conference, is our GP profession.

That our judgment, that we know how to judge the guideline, Our sense of responsibility for the patient in front of us, which is never the male 40-year-old person with one disease, and the relation we have with the patients.

These core values should weigh very heavily, but always in this interplay with with these guidelines.

So I hope I told you something about choosing wisely.

So for myself, I think it would be more wisely to come by bike next time.

Thank you.