Introduction
About 10 years ago some diseases were considered one disease, but now they are considered two different. Diagnostic tools for doctors today are what the state of science offers them today. About 10-20% of people are treated not for the disease that they have. It seems to you that 10-20-30% is a bit too much, but do not be surprised, the diagnoses can be erroneous.
And that’s why in medicine we would like to have such diagnostic tools to put as accurate diagnoses as possible so that later this person can be treated. Or nature will do its good work. In order to test how good the diagnostic method is, there is a very simple approach: to take in a controlled trial to test the application of this diagnostic technology, some super-duper magnetic tomograph – does it change the fate of patients or not? Is it useful for the outcome of the disease?
And it turns out that this simple method is very difficult to implement. And for the most part they do not study how much the healing method affects the outcome of the disease, but how accurate it is. And how accurate it is, you can learn only by comparing it with the most accurate diagnosis. And this is called the “measurement of diagnostic accuracy”.
Diagnostic accuracy is assessed from two sides: how well this method identifies patients – this is called sensitivity. For example, in order to learn from the smears from the nose, the flu in humans or not, there are express methods. Here these methods reveal the flu in 50-60% of those who really have the flu. This is low sensitivity.
The second characteristic is specificity. That is, how this method is good, so that there is no false positive result in healthy people. If there is a false positive result, then a person begins to be treated as a patient, but he does not have a disease. And good methods should be sufficiently high-sensitive and highly specific.
Unfortunately, there are no methods that are 100% accurate. And in this situation, the question should always arise: what diagnostic methods should we work with, what equipment should we buy? This requires a special field of activity, it is called “Evaluation of medical technology.” Unfortunately, we do not have this practice. It has been deployed in the most developed countries over the past 15 years, but we have not yet. Therefore, in our outpatient clinics and hospitals there are a lot of quite meaningless devices, from the point of view of modern science, to which, in addition, they spend money.
For example, today only on TV, once again demonstrated how in one of the Moscow polyclinics use computer and Internet tools to monitor the condition of patients at home. They give them devices to remove cardiograms and stuff. But there is no evidence that this kind of use of these things affects the outcome of the treatment! They are plausible quite – yes, a cardiogram with two hands is issued, but what is a cardiogram with two hands? It gives only the concept of a heart rhythm, but the heart rhythm can be determined by the patient at his own pace – rhythmically or not rhythmically beats. And money is spent on this!
Today we will pass to early diagnosis – it is denoted by the word “screening”, “screening”, which in a wide conversation is used little, probably because it is English. Screening – from the word screen – “select, sift”. Technology under this name was formed in the middle of the XX century, because it is by “sifting” the population that rare patients, for example, tuberculosis, can be identified. They are rare, but dangerous, because they infect others, and it was considered justified to do all radiographs and some tests so that all the rest remained healthy.
As always, I will begin the story on the next topic with the announcement that I received and receive money from state and non-state organizations that carry out medical work, but this lecture is not funded by anybody, no products are mentioned here, and this lecture I read at the invitation of the Polit.ru project.
The idea of identifying the disease as early as possible has a very good foundation, a legendary basis. It is believed that if you make a diagnosis earlier, it will be easier to cure. This idea is based on the idea that the disease is small at first, and then big. We all know that the bug can be easily crushed, but the elephant can not be crushed. Here, on the basis of such a household logic, the statement repeats that “we need early diagnosis, because it is with early diagnosis that the results of treatment will be the best.”
Hence the conclusion follows that the people will live longer. If the disease is cured early, then then it will not be necessary to spend money on the treatment of a serious illness – it will be detected at an early stage. From here it follows: if we treat all diseases early, then we will live forever?
Of course, I brought all these conversations in a somewhat ridiculous way, but that if the disease is identified earlier, it is easier to identify, you hear almost every day in one form or another. In fact, all this is either completely untrue, or almost completely untrue. For example, about reducing costs, if you carry out early diagnosis. There are good studies based on epidemiology, on the age distribution of the population, which show that the more prevention is done – good, effective – the more expensive becomes health care.
Why? Because life expectancy increases, the life expectancy in the terminal area proportionately increases, when a person is sick and it takes a lot of expenses, and in general everything turns out more expensively. Moreover, preventive interventions are not at all cheaper than medical ones, because, roughly speaking, 10 people are being treated, but in order to conduct a “sifting” and identify these 10 people, large expenses are needed for the initial stages of diagnosis.
We will return to this issue, but for now I want to draw your attention to the fact that there are many ways to take money from the population. Here you see an advertisement that is torn from the advertising magazine of Aeroflot. Tell me, which one of you was at the Ritz-Carlton Hotel? Oh, there are two people who were there, and those at work, probably. So, in this one of the most expensive hotels in Moscow, a doctor installed expensive equipment, and there are customers who come there and receive a magnetic resonance examination.
Do you offer a rental at this hotel? What does the doctor promise them? I highlighted in red: “Universities do not teach preventive medicine.” That is, prevention is prevention. It’s a lie, it’s taught, definitely taught! We in the law write that the main direction in health care is prevention. In general, the roots – in this same stupidity, that “the disease is easier to prevent than treat”, you heard, perhaps, such a phrase? In fact, it is still without proof that it is wrong, and the second, that this can also build a business. And such “preventive businesses” are arranged very well.
Let me note that just as there are private clinics that offer gynecological or urological services, there are clinics that parasitize on the idea of the utility of prevention. They offer surveys – they say, we will examine you, you will know your state of health, and then – just spit. In fact, problems with this are just beginning. Next we will see how problems start with the insane use of screening methods.
Of course, what I have said does not in any way indicate that effective preventive interventions do not exist. They are there, but for each preventive intervention evidence is needed. And this evidence must be good.
Last time I talked about diagnostic interventions. There are diagnostic methods that work, there are – which do not work. There are those that give a good balance between false positive and false negative results, and there are those that give results worse. With regard to prevention, there is exactly the same approach, the only thing that for prevention, good research is much more difficult to perform than research for treatment.
These tests are quite expensive, but fortunately, we have a reasonable approach, it is part of evidence-based medicine. What is it? “It is necessary to apply that which has been proved useful.” If it is proven that the method is harmful, then it should not be used. But, if there is no evidence of a method, then it DOES NOT have to be applied either! For it is possible that it will be harmful. Therefore, it is necessary to apply only that usefulness of what is proved.
And, I must say that the World Health Organization, which is spoken with great respect in our country, moreover, it was under the Soviet regime and now it also produces very valuable documents very often. One of them – in 1956, if I am not mistaken – in which the requirements to what screening methods should be formulated were formulated. What are the ways to examine people to identify diseases, and what are not.
Second: “the development of the disease must be well studied.” If we do not know if the disease is “small,” what is the point of early detection? First you need to understand how it develops. Let’s say that Ebola fever has no early stage that can be identified, it develops very quickly.
Fourth: “Early intervention should be more effective than in the later.” Well, of course, otherwise, why early identify?
“A diagnostic test should be available to identify the early stage.” Note: there must be more effective therapeutic intervention, and there must be a test that allows early detection. Because if we take those tests that doctors use to diagnose a disease, they can be useless for early diagnosis.
For example, an electrocardiogram – a method that came into being in the 1950s, in the clinic is necessary and very effective in individual cases. But, if we begin to examine healthy people with it, it will be a waste of time and paper. Because it is accurately shown that people can have very serious illnesses, including hearts that are not detected by electrocardiography, and vice versa, states that do not require any treatment at all. That is, there must be such an appropriate test.
Other criteria: “A test to identify the early stage of the disease should be acceptable.” For example, most people can easily spread a small piece of their feces on paper and send it by mail. This is acceptable for them. But the colonoscopy is not acceptable for everyone.
True, now modern colonoscopes have become thinner and lighter, but, nevertheless, people are less aware of the need for such prevention – examination of the colon through the anus. Not everyone agrees with this.
“Intervals between repeated examinations should be determined.” A person grows old, new diseases develop. How often should this be done? In a year or two years? Or every five years? The question is important, because costs are simply multiplied. And health systems are financially tense everywhere, not just in our country.
Please note that we are again looking at the level of common sense: “Additional diagnostic and therapeutic interventions that will be required as a result of screening should be provided.” We mentioned this problem last time: we have today the majority of people who need expensive help, for example, cancer, do not get this help because of a lack of funding. Queues, unavailable operations, unavailable, for example, radiotherapy in oncological dispensaries.
And in this situation, we suddenly – women! – the program of detection of breast cancer begins, and across the country tens and hundreds of thousands of women begin to come to dispensaries for further diagnosis after mammography.
Finally, there must be treatment tools available for those who have had the disease. If we do not have a way to treat this disease, why reveal it, but early? People are told: “Oh, yes you have a fatal illness, but we can not treat it” And this person would have lived 3-4 years before the clinical stage and felt healthy. A pretty result …
So pay attention: these fundamental problems with early examinations are well known. Already 60 years, as formulated approaches, with which every sober person will agree.
How would I like to see screening methods? It would be nice if it was a randomized controlled trial. We will approach this issue in more detail when we consider therapeutic intervention. As it was done, when the effectiveness of screening for breast cancer was studied.
This is a study scheme to study the effectiveness of screening for breast cancer using mammography. The study was conducted in Sweden. Randomly, there were selected regions where the mammography would be performed. And such regions where mammography will not be offered. In the region where mammography was actively offered, the number of women who made it increased many times over compared with the control areas.
Then in the regions of active detection, in the regions of screening, started once a year – every two years invite women and examine them. And in other regions they continued to work the way they used to work: they do not refuse mammography on necessity, but they do not actively invite them. During 5-6 years, collected data, and then compared. It was assumed that, where screening is active, more cases will be detected, but mortality will decrease, primarily from breast cancer.
Alas, it turned out that it almost does not work. Next, we turn to this phenomenon in more detail. I’m starting with breast cancer because this is the most famous example, and the second is a unique example, there are a lot of good tests done here, there is no more such well-studied example of screening.
Why such cumbersome research is absolutely necessary? If we carry out the investigation using a simplified scheme, then there will be so-called displacements that will shift our estimate. As we think, a chronic disease develops: first, some molecular processes take place at the intracellular level, which alter the metabolic or metabolic processes of DNA, that is, the subcellular processes of induction. They occur under the influence of ionizing radiation, or under the influence of infection by parasites or under the influence of infection, but gradually the cellular basis of chronic disease occurs.
Next – a period that is called latent, asymptomatic. The disease is already developing, but the person does not feel anything, lives, as if nothing had happened. At this time, the thickness of the vascular wall changes, lipid metabolism changes, depositions on the vascular wall are deposited – processes arise which, if strained, could be identified. But, while a person is healthy, he does not do it. And only then does the symptomatic period begin, during which normal medicine works.
Let’s pay attention that clinical diagnostics usually occurs not at the very beginning of the symptom period – because people do not treat their symptoms. Every normal person, when something starts hurting him somewhere, makes a normal conclusion: probably, soon will pass. And only when some time passes, people turn for help.
By the way, this means that if the availability of medical care is good, then early diagnosis in the symptomatic period increases. If we know that it is easy for us to get to the local doctor, if the district doctor can send us to the oncologist promptly, then this period is shortened. It is easier for us to go there, there are no queues.
But doctors want to go further in this latent period. And here they are faced with such a problem. Imagine that the disease occurs in 1990, and in 2002 a person dies. Usually a diagnosis is made to such a person in 1997. Doctors tensed and diagnosed earlier. In this case, the total life span of a person does not change, because there is nothing to treat. But doctors in their reports write that after the diagnosis the person began to live longer.
And all those who are engaged in screening, first of all, make this result. That is, if you look at the reports of those who conduct screening, then they have remarkable results. Even better results would be if all healthy people were called sick, then a significant life time would result after an early diagnosis. I emphasize that this is really proved in some trials, some studies, that such a mistake is being made.
Another hard-to-recognize error that you recognize. This is an error from the identification of long-running cases. In the figure, long and short cases are ranked. Imagine that we are here – again! – and surveyed all. And it turns out that all the long, slowly developing cases fall, and short cases end either before or after or here, and we from them catch mostly long-lasting cases.
So, if we take those cases that are identified in the screening, then they will be predominantly long, slow current cases. And this, too, is shown in benign studies that such an error occurs. 80 years ago there was no sustainable understanding that all diseases develop in different ways. But already in the fifties of the last century, with the example of lung cancer, it was shown that if you take the chest radiographs with an interval of a month, then in some people the tumor increases in size quickly, while in others it grows slowly. Accordingly, slow cases are better identified, and it seems that, by identifying these cases, we have done a good job, the tumor is slowly developing.
What should be the effective screening method? Here it is – screening for colon cancer. First, it is not external, but internal cancer, it is not visible, but you can still look there. Fortunately, he is also not badly treated. Therefore, today in all world programs for early detection of diseases, screening for colon cancer is mandatory. I must say with disappointment that our country here is far from an example. The screening program was absent in the 2006 medical examination program. The reason I do not know. But it was introduced only in the program of 2012.
How is the health of the population supposed to change if we start screening? The blue line is an increasing number of cases of colon cancer. Red dotted line – this is the people who are screened. What’s happening? First, the number of those who are diagnosed is increasing because they began to look. But further, since early cases have been revealed here, over the years, this quantity decreases, and this benefit becomes the benefit that one would like to receive.
And what happens at this time with mortality? Since these cases are early, they do not give rise to mortality. Rather, a small increase can be, but it is not noticeable with this method of measurement. From what? From the fact that those who are operated on, they have a chance to die on the operating table or in the postoperative immediate period. Those who are not operated, the chances for this are slightly less. But this harm of treatment is small, and people and doctors, accordingly, make the decision: let’s be operated on.
Here is this picture – if you listen to a lecture on medicine, you need to know it and learn to understand. This is the so-called “forest timetable”. It is used to generalize the data of several studies. Today, almost all important problems happen that there are 10-20-50 studies. No normal person can answer the question of whether this study helps or not, putting in front of these 50 studies. To do this, all the accumulated data should be systematized.
This is called systematic reviews. When they collect all the data on all the studies, they need to be generalized, this is done with the help of statistical technologies. If there were the same lethality in the screening group and in the absence of screening group – 60% there and 60% there, divide one by the other and get 1. And if there are 60 controls in the control and 30 at screening, . Accordingly, this point will leave here – 0.5.
This scale is a scale of relative risk or relative odds. We take the 2012 survey. This box shows that the reduction is, say, 20%. The second square is 20%. The third box in 2011 is almost 20%. Notice, the “antennae” appeared. What it is? This is the so-called “confidence interval”, a very important thing to the question of approximation.
If there are 10 people in one group and 10 people in another group, then our ratio of 0.5 will be very approximate. A total of 10 people. And this confidence interval indicates in what range the true estimate lies. And, when we say that there is a 20% reduction here, in reality it means “either 30%, or none at all.” You see, “antennae” go for a single. Very visual technology.
And, when the study is very small, like here – note the huge confidence interval. Nothing here can not say. And here is a larger study, the confidence interval is smaller. As a result, we get an estimate of 0.72, that is, almost 30% decrease in the likelihood of death from colorectal cancer, if screened. Still, 30% is very much! This is a good result, so we say that this is an effective technology.
Here’s another picture, it’s also about colorectal cancer. But there are three groups here that show us what happens if we just observe: a green twig is in the control group. Here, years – 5, 10, 15 years – imagine how difficult it is to conduct such a lengthy study! Here, generations of researchers changed, while monitoring this process.
And here on the ordinate – the mortality from colon cancer is summarized. Look – first for 5 years there is no difference. In 10 years the difference is already there. And we see that if you do screening more often, then the effect is greater. Hence, the rate of cancer development is comparable with the year. And there is the second schedule, here it is a question of the general death rate. It turns out that this process has no effect on the overall mortality rate.
I think this is an unexpected disappointment for you. There is a hypothesis, formulated 100 years ago, that a person with his severe chronic illnesses approaches, as it were, to the limit of his existence. And, if he does not die from this disease, then in the near future he will still die from the other. I understand that this is formulated quite sadly and rectilinearly, but I must formulate it in a straightforward way, so that it will be clear to everyone. This is the most plausible explanation why the prevention of this death leads to this result.
I must say that this is evident in studies of the prevention of cardiovascular diseases. There are absolutely delightful results for the prevention of cardiovascular diseases. But, when looking at the overall mortality, it turns out that there is almost no influence. By the way, here there is absolutely anecdotal consideration: all people would like to die of a heart attack, no one wants to die from colorectal cancer. Why prevent cardiovascular disease? It is distressing, but life, in general, does not promise great joy in the future.
