RK randomized clinical trials. Randomization procedure and code disclosure

Randomized controlled trial

Clinical trial (CT) is a prospective comparative study of the effectiveness of two or more interventions (therapeutic, preventive or diagnostic), which compares outcomes in groups of subjects that differ in the intervention used. In this case, the hypothesis about the effectiveness of the tested method (the effect of the intervention on the outcome), which arose before the study, is usually tested.

If there is a control group (comparison), we talk about controlled CT, and when groups are formed by randomization, we talk about randomized controlled trial (RCT, randomized controlled trial according to the classification of study types in MEDLINE).

Advantages - RCT results better reflect important differences in patient outcomes; systematic errors are least common; most objective for assessing the effectiveness and testing of interventions; the results of RCT performed strictly according to the research design are the most reliable.

Flaws - it takes a long time to conduct an RCT; they are expensive; not suitable for rare disease research; these studies have limited generalizability of the results (the ability to transfer the results to the population). This last limitation should not be exaggerated because other types of studies have even worse generalizability.

The study selects patients from a large population of people with the condition being studied. These patients are then randomly divided into two groups matched according to the main prognostic features. One group, called the experimental or treatment group, is given an intervention (such as a new drug) that is expected to be effective. The other group, called the control or comparison group, is subject to the same conditions as the first, except that its patients are not exposed to the intervention being studied. The reliability of clinical trials depends on the extent to which the compared groups were able to ensure the same distribution of all factors that determine prognosis, except for the therapeutic intervention being studied.

Sampling. Among the many reasons why patients with the disease being studied are not included in the study, the following three reasons are the main ones:

• 1) Patients do not meet the established inclusion criteria. This is the atypical nature of the disease, the presence of other diseases, a poor prognosis of the disease, and a high probability of non-compliance by the patient with the prescribed treatment. This limitation increases the reliability of the study: the possibility of outcomes not related to the treatment itself is reduced.
• 2) In cases of patient refusal to participate in experiments (clinical trials).
• 3) Patients who, in the early stages of the trial, showed an inability to strictly follow the prescribed treatment method are excluded. This will avoid wasted financial and therapeutic efforts and a decrease in the reliability of the study.

The following options for the structure of RCTs are distinguished.

Parallel- parallel (simultaneous) research in the active intervention and control groups is carried out independently of each other. This is the most common research structure.

Cross- a study conducted in one group of patients with a sequential change of treatment methods, separated by a “washout” period (to eliminate the effect of the previous intervention). Similar studies are carried out in patients with stable and usually chronic pathological conditions.

Steam room - a method of forming groups in a clinical trial, in which each participant in the main group corresponds to a participant in the control group, usually selected based on some common characteristics.

Sequential - a way of conducting a study in which the decision to stop is made when differences between groups are reached (usually the study is stopped at a predetermined time limit).

Factor protocol - The study is conducted in groups in which combinations of interventions are used. For example, in a 2x2 factorial protocol (for two treatments), four groups are formed, two of which use one of the treatments, the third uses neither of them, and the fourth uses both. The factor model is also used to assess the impact of different doses of one drug and combinations of drugs.

Adaptive - the enrollment of participants in the group receiving the worst treatment, according to cumulative estimates, decreases as the study progresses.

Design by Zelena - Participants assigned to the study intervention group are given the opportunity to withdraw from the intervention and enter the control group. Useful when studying interventions for which patients have strong preferences.

Compared to the parallel CI structure, other options are relatively complex both to perform and to understand their results and are usually used when the parallel structure seems inappropriate or impossible. The design of tests with these types of structures, as well as the analysis of the resulting data, requires consultation with a statistical specialist.

The test carried out is characterized by practical value, complexity and effectiveness. Treatment results should be reproducible and applicable in routine clinical practice. It is important to know whether the intervention being studied is sufficiently different from alternative treatments.

The value of the treatment method (drug) being studied can only be judged by comparing its results with the effect of other therapeutic measures, i.e. make comparisons between groups receiving different treatments. Or you can compare the effect of a treatment method with its absence. The latter method allows us to estimate the overall effect of health care, both related and unrelated to the intervention under study.

Placebo treatment. You can compare the effect of a study treatment (drug) with a placebo. A placebo is a dosage form that is indistinguishable from the test drug in appearance, color, taste and smell, but does not have a specific effect (for example, glucose tablets or injections of an isotonic solution. Placebo effect is a change in the patient’s condition (noted by him or her or the attending physician) associated with the fact of treatment, and not with the pharmacodynamic effect of the drug. The placebo effect is considered by researchers as a basal level for measuring specific treatment effects. It is necessary to distinguish between specific and non-specific effects of a treatment intervention for its objective evaluation in a clinical randomized trial sample.

Placebos in clinical drug trials are useful for the following purposes:

• 3) distinguishing between the actual pharmacodynamic and psychological effects of the drug;
• 4) the difference between the effects of the drug and spontaneous periodic remissions and the influence of other external factors;
• 5) avoid receiving false negative conclusions.

It is possible to compare the study treatment with usual care - this is acceptable in cases where the usual treatment has been shown to be effective.

blind method.

• 6) researchers who distribute patients into intervention groups do not know what treatment will be assigned to each subsequent patient;
• 7) patients should not know what kind of treatment they are receiving;
• 8) supervising doctors should not know what treatment (drug) is prescribed to the patient;

To study the specific therapeutic effect of an intervention (medicine), it is necessary to randomly assign patients to groups, i.e. by randomization. Randomization is the optimal method for choosing treatment, which avoids systematic error when dividing patients into groups. Randomization allows patients to be divided into groups with predominantly identical characteristics.

If trial participants know who is receiving what type of treatment, there is a possibility that their behavior will change, which could introduce bias. To reduce this effect, apply blind method. Blinding in clinical trials can be carried out at the following levels:

• 9) researchers who distribute patients into intervention groups do not know what treatment will be assigned to each subsequent patient;
• 10) patients should not know what kind of treatment they are receiving;
• 11) supervising doctors should not know what treatment (drug) is prescribed to the patient;

A “single-blind method” (the patient is not informed) or a “double-blind method” (both the patient and the researcher are not informed) are used. Thus, the “double-blind” method serves as a type of control to prevent the influence of bias on the results of the study.

The results of randomized controlled trials using a blinded (masked) design should be given precedence over any other information about the effects of treatment. However, such tests have limitations: they are expensive; there may not be a sufficient number of patients with the disease being studied; duration of the experiment; misunderstanding of doctors and patients about the need to conduct clinical trials and others. For many clinical questions, it is not always possible or practical to rely on the results of randomized clinical trials, so other evidence is also used.

In the 6th year at one of the departments, the teacher asked our group a question: “ On what basis are medications recommended for the treatment of a particular disease?" Some students suggested that drugs are chosen based on their mechanism of action, characteristics of the disease, etc. These are not entirely accurate answers. Nowadays, medications are chosen primarily based on their efficiency. And they do it with rigorous scientific methods. Today you will learn:

• Which study is cheaper - longitudinal or cross-sectional,
• that not only children like pacifiers,
• why blind treatment is considered the most valuable.

Modern treatments are based on positions evidence-based medicine (evidence based medicine). « Evidence-based medicine", also called " clinical epidemiology" Evidence-based medicine makes it possible to predict the course of a disease in a particular patient based on the course of many similar cases studied using strict scientific methods of mathematical statistics.

To draw any conclusions about the effectiveness or ineffectiveness of a medicine, studies are conducted. Before a drug is tested on real patients, it undergoes a series of experiments on living tissues, animals and healthy volunteers. I will talk in more detail about these stages separately in the material “How drugs are developed.” At the end, the effectiveness and safety of the drug is tested on a group of sick people, this trial is called clinical.

When preparing a clinical trial, scientists determine the population of subjects, selection and exclusion criteria, methods for analyzing the phenomenon being studied, and much more. All this together is called study design.

Types of clinical trials

Exists 3 types of clinical trials, which have their advantages and disadvantages:

• cross-sectional (single-stage) studies,
• longitudinal (prospective, longitudinal, cohort) studies,
• retrospective studies (“case-control”).

Now more details about each type.

1) Cross-sectional (single-stage) study.

This single examination of a group of patients. You can obtain, for example, statistics on the incidence and current course of the disease in the study group. Reminiscent of a photograph taken at a specific point in time. Cross-sectional studies are cheap, but they do not provide insight into disease dynamics.

Example: preventive examination of a shop doctor at an enterprise.

2) Longitudinal (prospective, longitudinal, cohort) study.

Terminology: from Lat. longitudinalis- longitudinal.

A longitudinal study is observation of a group of patients over a long period of time. Today, such studies are the most reliable (evidence-based) and therefore are carried out most often. However, they are expensive and are most often performed in several countries at the same time (i.e. they are international).

Why are longitudinal studies also called cohort studies? Cohort -

1. Tactical unit of the legion in Ancient Rome (from the 2nd century BC). There were 10 cohorts in the legion, with 360-600 people per cohort.
2. In a figurative sense, a tightly knit group of people.
3. In clinical epidemiology cohort- this is a group of individuals initially united by some common characteristic (for example: healthy individuals or patients at a certain stage of the disease) and observed over a certain period of time.

Single-arm study design diagram.

Prospective studies include simple and double-blind, open, etc., more on this below.

3) Retrospective studies (“case-control”).

These studies are carried out when it is necessary to connect risk factors from the past with the current condition of the patient. The simplest example: a patient has had a myocardial infarction, the local doctor flips through his card and thinks: “ Indeed, high cholesterol for several years does not end well. It will be necessary to prescribe statins to patients more often».

Retrospective studies are cheap, but have low evidence, because information from the past is not reliable (for example, the outpatient card could be filled out retroactively or without examining the patient).

Double-blind, randomized, multicenter, placebo-controlled study

As I mentioned above, the most conclusive are prospective (longitudinal) studies, which is why they are held most often. The most reliable of all prospective studies to date is double-blind randomized multicenter placebo-controlled trial. The name looks too scientific, but there is nothing complicated about it. I will explain the term word by word.

What's happened randomized trial? The word comes from English. randomize- arrange in random order; mix. Since the effectiveness of the drug being tested must be compared with something, each study has experimental group(it checks the required drug) and control group, or comparison group(patients in the control group are not given the test drug). Looking ahead, I will say that a study with a control group is called controlled.

Randomization in this case is the RANDOM distribution of patients into groups. It is extremely important that researchers, for their own selfish purposes, cannot collect milder patients into an experimental group, and more severe patients into a control group. There are special randomization methods to ensure that the differences between groups are not statistically significant. About the concept " reliability“I will also tell you further about evidence-based medicine.

What's happened blind and double-blind study? At single blind In the study, the patient does not know which group he was assigned to during randomization and what drug he is given, but the health worker knows this and may involuntarily or accidentally give away the secret. At double blind In the study, neither the doctor nor the patient is aware of what exactly a particular patient is receiving, so such a study is more objective.

Note. If for some reason it is not possible to use a placebo (for example, a doctor or patient can easily recognize a medicine by its effects, for example: MgSO4 when administered intravenously gives a short-term feeling of intense heat from the inside), carry out open study(both the doctor and the patient know which drug is prescribed). However, open-label research is much less reliable.

It is curious that of the total number of patients in the hospital placebo(dummy medicine; placebo imitates the drug, but does not contain the active substance) helps 25-35% , in cases of mental illness - up to 40%. If a patient receives a placebo and has a clear benefit, those patients may be excluded from the study.

Instead of a placebo, a drug can be used that they want to compare with the test drug. In turn, the drug being tested can be taken in one of 2 options:

• in parallel groups: i.e. one group takes the study drug, and the second (control) group takes a placebo or comparison drug.

Diagram of the parallel group study model.

• in a cross-sectional study: Each patient receives the test and control drugs in a certain sequence. There should be a free period between taking these drugs, designed to “eliminate” the consequences of taking the previous drug. This period is called " liquidation", or " laundering».

Diagram of the “crossover” research model.

What's happened controlled study? As I mentioned just above, this is a study in which there are 2 groups of patients: experimental group(receiving a new medicine or new treatment) and control group(NOT receiving it). However, there is a small problem. If you do not give the drug to patients in the control group, they will think that they are not being treated, and then they will become offended and depressed. The treatment results will definitely be worse. So the researchers give the control group a placebo - a dummy.

Types of control in evidence-based medicine:

1. Due to special fillers without the active substance, a placebo resembles the drug being tested in appearance and taste. This type of control is called placebo control (negative control).
2. If a patient taking a placebo is likely to suffer significant harm due to lack of treatment, the placebo is replaced with an effective comparator drug. This type of control is called active (positive). Active control is also used for advertising purposes to show that a new drug is superior in effectiveness to existing ones.

For completeness, I will also mention two rarer methods of control:

3. historical control, or control based on archival statistics. It is used when there are no effective treatments for the disease, and there is simply nothing to compare with. In this case, the treatment results are compared with the usual survival rate of such patients.

   Examples: some types of cancer treatment, organ transplant operations in the early stages of transplantation development.

4. initial state control. Patients are examined and treatment results are compared with the baseline state before the experimental treatment.

Multicenter called a study that is performed in several “centers” - clinics. Some diseases are quite rare (for example, certain types of cancer), and at a given time in one center it is difficult to find the required number of volunteer patients who meet the criteria for inclusion in the study. Typically, such studies are expensive and are carried out in several countries, being international. For example, many hospitals in Minsk also took part in them.

Control period

Every study should have control (introductory) period, during which the patient does not receive the test drug or a drug of a similar type of action, with the exception of life-saving drugs (for example, nitroglycerin for angina pectoris). In international trials, during this period it is usually prescribed placebo.

Study without control period and randomization(random assignment to groups) cannot be considered controlled, so its results are questionable.

Each study must clearly state inclusion and exclusion criteria patients from the study. The better they are thought out, the more reliable the results will be. For example, when studying the effectiveness of β-blockers as anti-ischemic drugs, we must exclude from the study patients taking other drugs with similar effects: nitrates and (or) trimetazidine.

Disadvantages of controlled randomized trials

1) Non-representativeness of the selected group, i.e. the failure of a given sample to accurately reflect the properties of the entire population. In other words, from this group of patients it is impossible to draw true conclusions about all patients with this pathology.

As I indicated just above, there are strict criteria for inclusion and exclusion of patients from the study. This is necessary to receive uniformity groups of patients that can be compared. Usually these are not the most severe patients, because... in seriously ill patients, the strict requirements of controlled studies cannot be met: the presence of a control period, placebo administration, exercise tests, etc.

For example, in a study RITA(1993) compared the results percutaneous transluminal coronary angioplasty(expansion of a narrowed artery by inflating a mini-balloon in its lumen) with coronary artery bypass grafting(creating a bypass for blood flow past the narrowed section of the artery). Due to the fact that the study included only 3% of patients of those subjected to coronary angiography, its results cannot be generalized to the remaining 97% of patients. The sample is not representative.

2) Conflict of interest.

When the manufacturing company invests a lot of money into clinical trials of his drug (i.e., actually pays for the work of researchers), it is hard to believe that the author will not make every effort to obtain a positive result.

For these reasons the results of single studies cannot be considered absolutely reliable.

PAGES OF THE NATIONAL SOCIETY FOR EVIDENCE-BASED PHARMACOTHERAPY

Randomized clinical trials and observational studies: relationship in the hierarchy of evidence of drug effectiveness

Sergey Yurievich Martsevich*, Natalya Petrovna Kutishenko

State Research Center for Preventive Medicine Russia, 101990, Moscow, Petroverigsky lane, 10, building 3

This article compares the role of randomized controlled trials (RCTs) and observational studies in assessing the effectiveness and safety of drugs in the field of cardiology. An unambiguous conclusion is made that RCTs are the basis of modern evidence-based medicine, and that there is no alternative to them. Observational studies conducted in accordance with modern regulations can provide information on the effectiveness of drugs only in the absence of RCT data.

Key words: randomized controlled trials, observational studies, comparison of information content in assessing the effectiveness of the drug.

For citation: Martsevich S.Yu., Kutishenko N.P. Randomized clinical trials and observational studies: relationships in the hierarchy of evidence for drug effectiveness. Rational Pharmacotherapy in Cardiology 201 6;1 2(5):567-573. DOI: 10.20996/1819-6446-2016-12-5-567-573

Randomised Clinical Trials and Observational Studies: the Ratio in the Hierarchy of Evidence of the Efficacy of Drugs

Sergey Yu. Martsevich*, Natalya P. Kutishenko

State Research Center for Preventive Medicine. Petroverigsky per. 1 0-3, Moscow, 1 01 990, Russia

The role of randomized controlled trials (RCTs) and observational studies in evaluation of the efficacy and safety of cardiology drugs is compared in the article. The clear conclusion is made that RCTs are the basis of modern evidence-based medicine, and that they have no alternative. Observational studies conducted in compliance with the modern regulations can be a source of information on the efficacy of drugs only in the absence of data from RCTs.

Keywords: randomized controlled trials, observational studies, comparison of informativeness in evaluation of drug efficacy.

For citation: Martsevich S.Yu., Kutishenko N.P. Randomised Clinical Trials and Observational Studies: the Ratio in the Hierarchy of Evidence of the Efficacy of Drugs. Rational Pharmacotherapy in Cardiology 201 6; 12(5):567-573. (In Russ). DOI: 10.20996/1819-6446-2016-12-5-567-573

Introduction

The need to demonstrate the positive effect of a drug or treatment on the outcome of a particular disease, as well as the safety of using this drug or treatment, is the basis of modern medicine. However, this was not always the case. For a long time, the main method of evidence in medicine was the so-called clinical experience. This was partly due to the fact that many of the drugs used (for example, nitroglycerin) had a quick and obvious effect that could be easily observed by the practitioner in everyday practice.

Generalization of clinical experience, especially if it was made by people authoritative in medicine, is often false.

Received / Received: 21.1 0.201 6 Accepted / Accepted for printing: 24.1 0.201 6

lived as the basis for mandatory treatment methods. For example, just 70 years ago, the classics of Russian therapy wrote about the treatment of myocardial infarction: “Complete rest and bed rest must be carried out strictly and for a long time. If the disease is severe, the patient should remain in bed for 2-3 months. Experience has shown that such long-term observance of rest reduces mortality from myocardial infarction....” Note that the authors who claim such a beneficial effect of rest on disease outcomes do not refer to any studies that have proven the effectiveness of such therapy.

Awareness of the fact that modern cardiovascular (and not only cardiovascular) diseases occur over a long period of time, passing through certain stages of development over many years, has led to the understanding that the treatment of such diseases requires drugs that are used over a long period of time.

neither. To evaluate the effect of these drugs, it became necessary to prove their effect on disease outcomes. Obviously, clinical experience was completely inapplicable for this purpose. There was a need to generalize clinical experience, process accumulated data, etc.

Observational studies

Clinical experience has been replaced by so-called observational studies. Their main feature is the lack of active controlled intervention on the part of the doctor. Main types of observational studies: cohort studies, case-control studies, cross-sectional studies. A description of the specifics of each of these types of studies is beyond the scope of this publication. Observational studies have played a role in drug research, but this role has been very limited. Although observational studies are very good at tracking disease outcomes, they do not always answer the question of what factors influenced this outcome. Attributing a positive effect of disease outcome to any drug that was actively prescribed in this type of research often led to erroneous conclusions, since the outcome of the disease was influenced by many factors and it was not always possible to isolate the effect of a particular drug among them.

A classic example is the use of antiarrhythmic drugs to treat acute myocardial infarction. Quite a long experience of their use convinced of their ability to eliminate arrhythmias. However, the subsequent controlled CAST study, while confirming the antiarrhythmic effect of these drugs, completely refuted the possibility of their positive effect on disease outcomes. Moreover, patients receiving antiarrhythmic drugs died significantly more often than patients who did not receive them. The results of this study completely changed clinical practice.

Randomized controlled trials - the basis of evidence-based medicine

Awareness of the limitations of assessing the effect of a drug in observational studies led to the understanding that it is necessary to introduce the principle of experimental research into medical clinical science, naturally, without violating the interests of the patient. The result was the emergence of so-called randomized controlled trials (RCTs). There is currently ongoing debate about

who and when conducted the first RCT in medicine. Another fact is more important: the introduction of RCTs was the first step in transforming clinical medicine from an art into a science and created a separate science, which received the now well-known name “Evidence-Based Medicine.”

A description of the basic principles of conducting RCTs is also not the purpose of this article; we will only note their main feature: randomization allows us to obtain two (or more) groups of patients identical in basic clinical characteristics, differing only in the fact of taking the drug under study. Modern evidence-based medicine considers RCTs as the highest degree of evidence. It has already been said that RCTs can refute many treatment principles that are considered obvious. A recent example of such an RCT is the just completed NORSTENT trial, which failed to demonstrate any benefit of drug-eluting stents over conventional stents on the long-term prognosis of coronary artery disease, although the benefits of modern stents were previously considered undeniable.

The characteristic features of RCTs noted above have led to the fact that at present, when introducing new drugs into clinical practice, making decisions about their clinical registration and prescribing rules, the results of RCTs performed with these drugs are primarily taken as a basis.

Modern clinical guidelines have for many years used a so-called rating system, which allows each clinical decision to be assigned a class of recommendations with a certain level of evidence (evidence). RCTs in this system occupy the highest levels - A or B, depending on the number of RCTs conducted and the strength of the data obtained in them. On the other hand, observational studies and, first of all, registries are assigned a more modest role. It should be noted that registers in this rating system did not appear immediately, and the results obtained in them are classified as a low level of evidence - level C. In contrast to the European and domestic rating system, several years ago, when creating the ACC/AHA recommendations (American College of Cardiology/ American Heart Association) began to use a higher degree of detail in the evidence system, while the status of registers, qualitatively performed and analyzed using modern statistical methods and approaches, increased to level B (B-NR, NR - nonrandomized, non-randomized studies).

Of course, RCTs are not without certain disadvantages or limitations, the main one of which is considered to be the high selectivity of patients selected for participation.

sty in them. In addition, RCTs have a relatively short follow-up period, which many consider insufficient to fully identify both the positive and negative properties of the drug. It should immediately be noted that most of the limitations of RCTs given later in the article are characteristic not so much of RCTs as such, but of individual RCTs and largely depend on the purpose set in the RCT, its protocol, criteria for including patients and methods for assessing the main and side effects of the drugs being studied. . Due to the presence of such limitations, a tendency has recently emerged to contrast RCTs with certain types of observational studies, which, as some authors believe, do not have a number of disadvantages of RCTs.

As already noted, the main limitation of RCTs is the selectivity of the patients included in them. Some authors argue that such strict selection of patients in RCTs leads to the inclusion of “refined” patients, whom the doctor rarely encounters in real practice. It should be noted that this disadvantage is relative. Of course, expanding the exclusion criteria for RCTs makes the results more predictable and better interpretable, and this technique, unfortunately, is increasingly used in recent years. However, many large RCTs (eg, the ISIS-4 study, which included more than 58,000 patients with suspected acute myocardial infarction) had few exclusion criteria and, therefore, their results are applicable to a much wider patient population than the results of those RCTs. where exclusion criteria are numerous. The main thing to remember is one of the basic principles of evidence-based medicine: the results of a particular RCT are applicable only to the same patients who participated in it. Transferring the results of an RCT to a broader population of patients (ie, to patients who were not included in this RCT) is a big mistake.

Another disadvantage of RCTs is the limited follow-up period. Indeed, the duration of some RCTs is short. For example, in the MERIT-HF study, which studied the effect of beta-blockers on the outcomes of severe heart failure, the follow-up period was 1 year. This approach is often explained by the desire to quickly obtain results for a specific drug. However, there are RCTs where the observation period for patients is much longer. An example is the ATLAS study, which included approximately the same patients as the MERIT-HF study, but whose follow-up period was already 39-58 months.

It is also believed that RCTs do not fully identify the side effects of a drug, and some

side effects can only be detected during post-marketing studies of the drug. However, this opinion is erroneous. A drug such as cerivastatin (from the group of statins) is often cited as an example. The safety of cerivastatin in RCTs did not differ from the safety of other drugs in this group, but with widespread clinical use, it was noted that the potentially fatal complication of rhabdomyolysis occurred with cerivastatin significantly more often than with other statins. However, the evidence base for this drug was extremely small; as it turned out, only 2 RCTs were conducted with it, which included only about 1000 patients, and the follow-up period was very short.

It should be noted that not all RCTs aim to identify all side effects of a drug (although all side effects in RCTs are strictly recorded in accordance with GCP rules). There are also opposite examples when RCTs include, along with the primary efficacy endpoint, so-called safety endpoints. Examples include large RCTs that have been conducted with new oral anticoagulants.

Most experts in evidence-based medicine believe that drug-specific adverse events (AEs) can only be identified in RCTs, provided that the protocols for such studies set such a task. This is explained by the fact that only in RCTs can one judge with a high degree of certainty the connection between an identified AE and the drug taken. An example is the same studies with new oral anticoagulants.

It is impossible not to mention another important factor that must be taken into account when assessing the results of RCTs and their practical significance: different RCTs have very different quality. This concerns the correct choice of the number of included patients, the determination of primary and secondary end points, the general study protocol, the choice of comparator drug, etc. Therefore, the evidentiary value of different RCTs can vary greatly; often obvious methodological errors in a particular RCT lead to the fact that the conclusions made by the researchers , become unconvincing.

In table Table 1 shows the main differences between RCTs and observational studies in terms of the ability to assess the effectiveness of a drug.

Are randomized controlled trials always necessary to prove the effectiveness of a drug?

Despite the obvious fact that RCTs are the “gold standard” of evidence-based medicine, it is not always necessary to prove the effect of a drug.

Table 1. Comparison of randomized controlled trials and observational studies to evaluate the effect of drugs Table 1. Comparison of RCTs and observational studies to evaluate the effect of drugs

Parameter Randomized clinical trials Observational studies

Protocol rigor Type of experimental study. Randomization, as a rule, allows one to minimize the influence of confounding factors and isolate the effect of the intervention (drug). The high influence of confounding factors does not always allow one to isolate the effect of the drug (even after using “pseudo-randomization”)

Representativeness of the sample The patients included are not always typical of routine clinical practice, but this depends on the strictness of the inclusion/exclusion criteria. The sample is usually much more representative, but this depends on the selection of the patient cohort. The use of registers increases the representativeness of the sample

Monitoring adherence to treatment High. It is possible to use direct methods Low, if additional methods are not used (questionnaires)

Evaluation of the results of long-term treatment Limited by the timing of RCTs; as a rule, the duration of treatment is shorter than in observational studies. The duration of observation is, in principle, unlimited. Monitoring of patient dropout is necessary (not always easy to implement)

Assessment of side effects is much more accurate due to closer monitoring of patients and the ability to establish a relationship with the drug being studied. It is not always possible to identify rare side effects and side effects that occur during long-term treatment. Sometimes it is possible to identify side effects that did not appear in RCTs (due to a longer observation period and a wider representation of patients with comorbid pathologies)

Assessment of drug interactions Not always possible due to strict criteria for concomitant therapy Provides a wide opportunity for study, but it is not always easy to establish cause-and-effect relationships

Difficulty of carrying out High cost of carrying out. Requires lengthy preparation. Relatively inexpensive. Allows you to quickly get results.

Evaluation of drug effects in different subgroups Possible if planned in advance Possible, but results may be invalid due to confounding factors

they are being carried out. This primarily applies to drugs used to treat widespread diseases that have a rapid and distinct effect. Modern medicine knows a number of drugs with which RCTs have never been conducted, but whose effectiveness no one doubts (belong to class I recommendations). One example is the antibiotic penicillin, the use of which in the early 40s. made it possible to reduce the extremely high mortality rate in lobar pneumonia by more than 2 times. In the field of cardiology, such examples are practically absent; one of the few examples in this area is the defibrillation method, to prove the effectiveness of which the question of the need for an RCT has never been raised. Drugs used for the prevention (both primary and secondary) of cardiovascular diseases require long-term use, their effect is not so obvious and is not always consistent in all patients. Their effectiveness can be proven only by comparing the probabilities of adverse events in the main and control groups, for which an RCT is necessary. An example is the study of oral anticoagulants in the prevention of cerebral stroke in patients with atrial fibrillation.

Situations are acceptable when conducting an RCT is, in principle, impossible for a variety of reasons (Table 2); in such cases, there is no other choice but to turn to data from large registers and try to evaluate with their help the real effectiveness of the drug (including comparative). An example of this approach is an attempt to evaluate the current role of beta-blockers in the treatment of coronary artery disease, especially in patients who have suffered acute myocardial infarction. The relevance of this approach is dictated by the fact that the main studies performed with beta-blockers were carried out quite a long time ago, when there were no ACE inhibitors, statins, and thrombolysis and invasive methods of revascularization were not used. A natural question arises whether beta-blockers in modern conditions affect disease outcomes in the same way as they did in RCTs conducted 30-40 years ago. From a scientific point of view, this issue could only be resolved by conducting a new RCT, but this is impossible, primarily for ethical reasons.

Improving the methodology for conducting observational studies

Modern observational studies, first of all, registries try to take full advantage of their positive properties, first of all,

the possibility of including a huge, practically unlimited number of patients corresponding to real clinical practice and an almost equally unlimited period of observation. Neither one nor the other, however, eliminates observational studies from their main drawback - the presence of so-called biasing factors. Recently, a number of techniques have appeared (logistic regression, propensity score matching, the so-called “propensity score matching” technique for identifying patients who have the same indications for prescribing a particular drug). This allows the formation of groups within cohorts of patients included in observational studies that do not differ in baseline indicators, but differ in whether the drug of interest to the researchers was prescribed or not. The use of such methods makes it possible to perform the so-called “pseudorandomization” and, as it were, imitate an RCT. This has even led some researchers to conclude that these types of observational studies could replace RCTs, or at least reduce their role in evidence-based medicine.

However, in addition to certain technical imperfections, all these methods have one very significant limitation: there is never any certainty that they were able to take into account all the factors influencing the outcome of the disease, and, accordingly, this does not allow us to draw conclusions that the achieved result was obtained due to the action of the drug of interest drug.

Can questions unanswered by randomized controlled trials be addressed by observational studies?

A good example of such a formulation of the question in modern cardiology is the debate about which of the three new oral anticoagulants that have appeared recently - dabigatran, rivaroxaban or apixaban (a fourth drug, edoxaban, has recently been added to them) is more effective and safer. Each of these drugs was studied in a large RCT compared with the standard anticoagulant warfarin. Each of these drugs showed a positive effect on the incidence of the primary endpoint, which was almost the same in all of these RCTs. However, there have been no direct comparisons between new oral anticoagulants in RCTs (such studies are unlikely to ever be conducted for purely ethical reasons). Therefore, it is in principle impossible to answer the question of which of the 3 new oral anticoagulants is more effective and safe from the standpoint of evidence-based medicine.

Attempts are being made to do this through observational studies, mainly large registries. For example, one such study found that dabigatran and apixaban were the most effective in reducing the risk of death and bleeding compared with warfarin. From our point of view, such attempts are obviously doomed to failure due to the inability to fully take into account all the so-called interfering factors (Table 1). In simpler terms, we can say that in real clinical practice, each doctor has his own preferences in choosing each of these drugs (the most difficult factor to take into account); the regimen for prescribing them according to official instructions differs (frequency of administration, taking into account the degree of renal dysfunction). Therefore, in an observational study it is extremely difficult to obtain completely comparable groups of patients (even using special statistical approaches) that differ only in which of the new oral anticoagulants was prescribed. Accordingly, comparison of these drugs in terms of their effect on long-term disease outcomes in an observational study will never be completely correct. By the way, the authors of such studies, as a rule, clearly record these facts, recognizing the limitations of such analyzes.

What is the current role of observational studies?

First of all, it is necessary to answer the question of whether all observational studies meet certain quality standards (there is one for them too). First of all, we mean the representativeness of the sample included in such studies. The most representative sample can be ensured by modern registers, which also have certain requirements, but their description is beyond the scope of this publication. Let us only note that recently there has been a tendency to call databases registers, which are increasingly appearing in various fields of medicine. In this regard, it should be emphasized that a register and a database are not the same thing. A registry is understood as “an organized system that uses observational research methods to collect uniform data (clinical, etc.) and that serves a predetermined scientific, clinical, or organizational-methodological purpose.” Therefore, when planning, for example, to study a drug in more detail within the framework of the registry, they plan in advance (within the framework of real clinical practice) to ensure monitoring of its clinical effect, safety, adherence to its use (for this it is possible to use

Table 2: Situations when conducting observational study to evaluate the effect of the drug is possible / necessary in the absence of randomized controlled trials Table 2. Situations when conducting observational study to evaluate the effect of the drug is possible / necessary in the absence of RCT results

Situation Example Comment

When a newly created drug is clinically in demand and has a pronounced, distinct and rapid effect. Use of penicillin for the treatment of lobar pneumonia. It allowed to reduce mortality by 2 or more times. Conducting a subsequent RCT seemed impractical and unethical. Observational studies were needed to assess the safety of the drug

When it is basically impossible to conduct an RCT The use of cardiovascular drugs in pregnant women The need for information on effectiveness and safety is very high. It is necessary to evaluate the effectiveness and safety within the framework of observational studies (registries)

When the results of previous RCTs are outdated Use of beta-blockers in patients who have had myocardial infarction Basic therapy for myocardial infarction has changed significantly over 30 years (thrombolysis, ACE inhibitors, angioplasty). Conducting new RCTs with beta-blockers is unethical. Output: assessment of the effect of beta-blockers in modern conditions within the framework of registries

When a hypothesis arises about new indications for the use of an already registered drug and a widely used drug. Use of ursodeoxycholic acid to enhance the effect of statins. The RAKURS study was assessed using the propensity score method. The result requires confirmation using an RCT

special questionnaires). Databases do not provide such an opportunity; adherence to therapy is usually assessed in them using written prescriptions; this approach can create a picture of adherence that is very far from the real one.

The main tasks of modern registers as the most advanced form of observational research are seen as follows. Firstly, this is obtaining a so-called “portrait” of a typical patient with a particular disease (or a combination of them), i.e. basic characteristics of the patient, including demographic, socio-economic and clinical. The characteristics of patients in different countries and different regions of the same country may differ significantly. By comparing the “portrait” of a patient obtained in a particular registry with the “portrait” of a patient who participated in a particular RCT, we can draw a conclusion to what extent real patients correspond to the patients who participated in a particular RCT, and, accordingly, conclude that how applicable the results of RCTs are to patients included in the registry. For example, after analyzing the available Russian registers that included patients with atrial fibrillation, it was concluded that, on average, Russian patients with atrial fibrillation have a more severe course of the disease than patients included in studies comparing new oral anticoagulants and warfarin. The patients included in the ROCKET-AF study, which compared warfarin with rivaroxaban, turned out to be the closest in characteristics to Russian patients.

Second, registries provide invaluable information regarding adherence to therapy. This is ka-

This concerns both the adherence of doctors to compliance with modern guidelines and the adherence of patients to the treatment prescribed by doctors.

Thirdly, registries make it possible to track disease outcomes over an unlimited period of time. Naturally, it is possible to assess the influence of various factors on disease outcomes, including medications, as discussed above. However, with such an analysis, a number of methodological problems arise (often insurmountable), especially when attempts are made to evaluate not the role of any one drug, but to compare several drugs with each other.

As discussed earlier, in some situations it is not possible to conduct an RCT, in which case registries must be used to evaluate the effectiveness of a drug. A good example of such an analysis is an attempt to assess the role of beta-blockers for the treatment of coronary artery disease in modern conditions. Bangalore S. et al conducted pseudo-randomization within the REACH registry, simulated a randomized trial and came to the conclusion that the role of beta-blockers in modern conditions has indeed become less significant. We note, however, that the authors of modern clinical guidelines did not react in any way to the results of this (and several similar) analyzes, not considering them convincing enough to reconsider the role of beta-blockers in various forms of coronary artery disease.

Fourthly, registers provide the opportunity to conduct so-called pharmacoeconomic studies. Along with electronic documentation, registers are becoming one of the most important sources of information for performing clinical and economic

Russian studies that allow not only to assess the effectiveness and safety of certain interventions, but also to calculate the costs of their use. Obviously, carrying out such calculations allows us to develop rational tactics for managing patients, taking into account the economic aspects of the examination and treatment.

Conclusion

To summarize, we note that modern RCTs are the basis of modern evidence-based medicine; today there is no alternative to them in terms of assessing the effect of the drug. The absence of RCTs on any issue and their replacement with data from observational studies sharply reduces the degree of evidence

importance of a particular fact, which is reflected in clinical recommendations in the form of a lower level of evidence and class of recommendations.

Observational studies, when conducted in accordance with established rules, play a huge role in drug evaluation, but this role is fundamentally different from the role of RCTs.

Antitumor drugs vary, each of them is carried out for specific purposes and is selected according to the necessary parameters for studying the drug. Currently, the following types of clinical trials are distinguished:

Open and blinded clinical trial

A clinical trial can be open and blind. Open study- this is when both the doctor and his patient know which drug is being studied. Blind study divided into single-blind, double-blind and completely blind.

• Single blind study- this is when one party does not know which drug is being studied.
• Double blind study And completely blind study is when two or more parties do not have information regarding the study drug.

Pilot clinical trial is carried out to obtain preliminary data important for planning further stages of the study. In simple language one could call it “sighting”. With the help of a pilot study, the possibility of conducting a study on a larger number of subjects is determined, and the necessary capacity and financial costs for a future study are calculated.

Controlled clinical trial is a comparative study in which a new (investigational) drug, the effectiveness and safety of which has not yet been fully studied, is compared with a standard method of treatment, that is, a drug that has already been studied and entered the market.

Patients in the first group receive therapy with the study drug, patients in the second receive standard therapy (this group is called control, hence the name of the type of research). The comparator drug can be either standard therapy or placebo.

Uncontrolled clinical trial is a study in which there is no group of subjects taking a comparator drug. Typically, this type of clinical research is carried out for drugs with already proven effectiveness and safety.

Randomized clinical trial is a study in which patients are randomly assigned to several groups (by treatment or drug regimen) and have equal opportunity to receive the study drug or a control drug (comparator or placebo). IN non-randomized study There is no randomization procedure; therefore, patients are not divided into separate groups.

Parallel and crossover clinical trials

Parallel clinical trials are studies in which subjects in different groups receive either only the drug being studied or only a comparison drug. A parallel study compares several groups of subjects, one of which receives the study drug, and the other group is a control. Some parallel studies compare different treatments without including a control group.

Crossover clinical studies are studies in which each patient receives both drugs being compared, in a random order.

Prospective and retrospective clinical study

Prospective clinical study– this is observation of a group of patients for a long time, until the onset of outcome (a clinically significant event that serves as the object of interest of the researcher - remission, response to treatment, relapse, death). Such research is the most reliable and therefore is carried out most often, and in different countries at the same time, in other words, it is international.

Unlike a prospective study, retrospective clinical study On the contrary, the outcomes of previously conducted clinical trials are studied, i.e. outcomes occur before the study begins.

Single-center and multicenter clinical trial

If a clinical trial takes place at a single research center, it is called single-center, and if based on several, then multicenter. If the study is carried out in several countries (as a rule, the centers are located in different countries), it is called international.

Cohort clinical trial is a study in which a selected group (cohort) of participants is observed over a period of time. At the end of this time, the results of the study are compared among subjects in different subgroups of this cohort. Based on these results, a conclusion is drawn.

In a prospective cohort clinical study, subjects are grouped in the present time and followed in the future. In a retrospective cohort clinical study, groups of subjects are selected based on archival data and their results are tracked to the present.

What type of clinical trial will be most reliable?

Recently, pharmaceutical companies have been required to conduct clinical studies in which results are obtained. the most reliable data. Most often it satisfies these requirements prospective, double-blind, randomized, multicenter, placebo-controlled study. This means that:

• Prospective– observation will be carried out for a long time;
• Randomized– patients were randomly assigned to groups (this is usually done by a special computer program, so that eventually the differences between the groups become insignificant, that is, statistically unreliable);
• Double blind- neither the doctor nor the patient knows which group the patient fell into during randomization, therefore such a study is as objective as possible;
• Multicenter– performed in several institutions at once. Some tumor types are extremely rare (for example, the presence of ALK mutations in non-small cell lung cancer), so it is difficult to find the necessary number of patients meeting the inclusion criteria in a single center in a single center. Therefore, such clinical studies are carried out in several research centers at once, and, as a rule, in several countries at the same time and are called international;
• Placebo-controlled– participants are divided into two groups, some receive the study drug, others receive a placebo;

Randomized trials are a precise method for identifying the cause-and-effect relationship between therapy and disease outcome, as well as the effectiveness of treatment.

General information

In the modern world, there are many medicines used to treat various diseases. According to advertising from drug manufacturers, they are all effective and have virtually no contraindications or adverse reactions. However, their levels of proven effectiveness vary. New drugs undergo numerous tests before appearing in the pharmacy chain. It should be noted that about 90% of them are rejected at the stages of clinical trials.

Evidence-based medicine

Since ancient times, various medicines have been used to treat ailments. And only starting from the nineteenth century they began to think about the effectiveness of drug therapy and the possibility of using scientifically based mathematical approaches in assessing the quality of treatment. Evidence-based medicine - this concept was first voiced by epidemiologists at a Canadian university when developing a training program for practical medicine. Doctor D.L. Sackett gave an official definition of this term.

Evidence-based medicine is the accurate, conscious, common sense use of the results of clinical research confirming the safety and effectiveness of any treatment methods that are best at a particular time. Clinical studies may refute therapeutic treatments that have been used successfully in the past with good results. They also formulate different approaches to treating patients.

Here is one example. During testing of antiviral drugs, it was found that they reduce the risk of developing pneumonia as a complication after influenza. Therefore, I prepared recommendations that included antiviral drugs for the treatment of this disease. In the modern world, when choosing therapy for treating patients, medical workers rely on evidence-based medicine and try to use new medications. Evidence-based medicine makes it possible to make a prognosis of the course of a disease in a particular individual based on similar cases that have been studied previously.

Placebo - what is it?

This is a substance that is similar in appearance to the test drug, but does not have its properties and does not cause harm to humans when taken. A drug is considered effective if its benefits, according to statistics, differ from those of a placebo drug.

In this case, one important condition must also be met, namely, the doctor and the patient do not have the right to know what exactly the patient is taking. This technique is called double-blind. In this case, the subjective opinion of medical workers about the therapy being performed and the indirect influence on the individual are excluded. There is also a triple blind method. In this case, the person who monitors the results of the study does not have information about how the patient groups, including the placebo groups, were selected.

Research

Trials are conducted on individuals to evaluate the effectiveness and safety of a new drug or to expand the indications of a drug already on the market. Clinical research is an integral stage in the development of new drugs; it precedes its registration. An experimental study examines a drug to obtain information about its safety and effectiveness. And based on the data received, the authorized body of the healthcare system makes a decision on registration of the drug or refusal.

Medicines that do not pass these tests cannot obtain registration and enter the pharmaceutical market. According to information from the Association of American Developers, as well as manufacturers of drugs for medical use, of approximately ten thousand drugs that were being developed, only 250 reached the stage of preclinical testing. Only five enter the next stage, i.e. clinical trials, of which only one is subsequently used by practicing doctors. Clinical studies provide knowledge both to medical professionals in terms of more accurate prescriptions, and to patients in terms of informing them about possible adverse reactions and contraindications.

Stages of clinical trials

There are several phases of experimental research.

The first takes about a year. During this period, the following indicators are examined: distribution, metabolism, absorption, excretion, dosage level, and the most convenient dosage form is selected. Healthy volunteers assist in this trial.

In the case of conducting research on drugs with high toxicity, people with the corresponding pathology are involved. Tests in such situations are carried out in specialized healthcare institutions that have the necessary equipment and trained medical personnel. The participation of volunteers, and they are usually required from 20 to 30 people, is financially encouraged in research.

Second, during this period the dosage regimen and dosage of the drug for the subsequent phase are determined. A group of volunteers ranges from 100 to 500 people.

The third is a randomized trial in which a large number of people (three thousand or more) take part. In this phase, the data obtained in the second stage on the effectiveness and safety of the drug in a certain group are confirmed or refuted. In addition, the dependence of the effect of the drug on the dose taken is studied and compared, as well as taking the drug at different stages of the disease or simultaneous use with other drugs.

Fourth - at this stage, clinical trials are carried out, necessary to obtain additional information about the effect of the drug, including in order to detect rare, but very dangerous side effects during long-term use in a large group of volunteers.

Necessary Requirements

For the reliability of scientific research when testing drugs, certain rules must be followed, thanks to which false results are minimal.

1. Large sample. The more patients studied, the lower the error.
2. Statistical processing of the obtained data. It is carried out taking into account the parameters being studied and the sample size. In this case, the error should not exceed seven percent.
3. Control or placebo groups. These are patients receiving a placebo drug instead of the study drug or standard of care treatment.

Types of clinical trials

There are several types, each of which has advantages and disadvantages.

• Single-stage or transverse. A group of patients is examined once. The cost of this type of research is low. Using it, it is possible to evaluate the incidence statistics and the course of the disease at a certain point in the study group. The dynamics of the disease cannot be revealed.
• Longitudinal or cohort. This type of research is considered the most evidence-based and is carried out frequently. A group of volunteers is monitored over a long period of time. The cost of its implementation is high, it is carried out simultaneously in several countries.
• Retrospective. Cheap type of testing, low, therefore unreliable. Used to identify risk factors. Data from previous studies are studied.

Randomization or random assignment

This is another rule that must be followed. Patients participating in the study are combined into spontaneous groups regardless of age and gender, i.e., candidates are randomly selected, which eliminates the influence of these factors on the results of the study.

The name “gold standard” is given to placebo-controlled, randomized studies using a double- or triple-blind method. Thanks to such tests, the information obtained is the most reliable. Unfortunately, due to the rather high cost and complexity, they are rarely carried out. In accordance with the basic tenets of evidence-based medicine, to make decisions on treatment tactics for a patient, physicians must be guided by the standardized international classification of clinical trials.

Difficulties

The difficulty of recruiting volunteers is considered one of the serious and difficult problems faced by research professionals. In general, about six percent of patients can be included in the single disease group.

Thus, the results obtained apply only to patients whose characteristics are identical to those studied in the groups. Therefore, it is not possible to recommend them for use in other conditions without obtaining a new test result. In addition, it should be taken into account that randomized trials do not at all mean the exclusion of erroneous results in the analysis.

Types of controlled tests

They may be:

• Single-center, when research is carried out in one healthcare institution. Difficulties - it is difficult to make a sample for all the studied characteristics in a short time.
• Multicenter. Several medical organizations are involved in the process, and all work according to the same protocol.
• Open. After randomization, the volunteer and the doctor have information about the type of treatment.
• Blind. The doctor will learn about the therapy after randomization, but the subject will not know about it (this issue is discussed in advance and the citizen’s voluntary consent to his participation in the research process is obtained).
• Double blind. In this case, neither the volunteer nor the doctor knows what type of intervention a particular individual will have.
• Triple blind. This type of test implies that the doctor, the subject and the researcher himself, who processes the results, do not have information about the type of intervention.

Disadvantages of randomized controlled trials

Due to high material costs and a long period:

• trials are carried out over a short period of time on a small group of volunteers or most studies are not carried out at all;
• in connection with the payment for tests by pharmaceutical companies, research institutes, universities, their direction is also indicated;
• Indirect evaluation criteria are used instead of clinical ones.

Systematic errors occur for the following reasons:

• inclusion in the group only of those volunteers who will give a predictable result when taking the drug;
• imperfect randomization;
• informing researchers about the placement of patients in specific groups, i.e., the blind method is not observed.

Benefits of randomized controlled trials

1. The effectiveness of the drug is assessed in comparison with a placebo drug in certain groups, for example, in men aged 40 to 50 years.
2. Accumulation of information after the study.
3. The goal may not be the ability to confirm one’s own hypothesis, but an attempt to falsify it.
4. Elimination of errors, since comparisons are being made in other identical groups.
5. The ability to combine the results obtained from several studies (meta-analysis).

Randomized trials are controlled trials that are double- or triple-blind and are classified as Class 1. The materials and information obtained as a result, as well as the meta-analysis performed, are used in the practice of medical workers as the most reliable source of information.

Conclusion

To introduce evidence-based research into medical practice, it is necessary to clearly describe the groups of volunteers on whom the effect of the drug for the treatment of a specific pathology was studied. Clearly specify the selection criteria and reasons for excluding patients from the trial, as well as evaluate the results using means available in practical medicine.