Q: How long does the state registration of a new drug take?

State registration of pharmaceuticals is carried out by a relevant authorized federal executive authority within the period less than two hundred and ten business days since the date of receipt of application on the pharmaceutical state registration. The aforesaid period includes time necessary for repeated carrying out of expertise of pharmaceuticals and (or) ethical review in accordance with the Article 25 of the present Federal law. The period of state registration of pharmaceuticals is calculated since the date of receipt of the application on the pharmaceutical state registration by a relevant authorized federal executive authority with enclosure of necessary documents till the date of issue of registration certificate of pharmaceuticals. The period of holding the clinical research of the pharmaceutical is not taken into consideration when calculating the period of its state registration.

Q: What are the types of randomized research blindness?

Open randomized research — either the patient or the doctor will learn the type of treatment to be applied immediately after the randomization. This type of research is generally used in cases when it is very difficult or ethically unreasonable to form complete simulation of the studied interference (for example comparison of conservative treatment of cerebral edema with surgical decompression).

Simple blind randomized research — the patient is not told about the type of the applied treatment and this aspect is discussed with the patient in advance in the course of receipt of his informed consent for the research. The doctor learns the treatment variant to be applied for the patient after the randomization procedure.

Double blind randomized research — neither the doctor nor the patient knows which type of interference will be applied. In cases when the studied drug needs to be prepared for extempore intake on the clinical basis (for example it needs to be diluted by some amount of physiological salt solution), at least one member of the research team (directly engaged in preparation of the solution) will clearly know what the patient will take. As a rule in these cases two independent monitors of clinical research are delegated for the clinical basis, one of them inspects documentation of only “blinded” research workers, and the second one — of “unblinded” ones. It is very important to understand that keeping patients’ allocation to the treatment groups secret is fundamental for unbiased assessment of the treatment results by the very patient and his observer, “blinded” research worker, that is why the “unblinded” personnel must never share available information with his/her colleagues.

Triple blind randomized research — the patient, the doctor and the statistician processing the research results do not know the type of interference in every compared group. During key research an independent committee for data assessment is often formed, and it holds interim performance and safety analyses before complete placement and closing of data base. Of course it is important for the committee members to receive already decoded data, that means to know which treatment every patient received. In case of triple blind randomized research an independent “unblinded” statistician or a group of statisticians is used for preparation of report for the committee.

Q: What are the main practical ways of carrying out randomization during the clinical research?

Randomization can be carried out directly on the clinical base or centrally. In the first case the method of envelopes is often applied. Herewith the sponsor (statistics department or the contractor responsible for the randomization procedure) provides to the research worker series of numbered and sealed envelopes, and in every such envelope one of the compared treatment types shall be specified. When including the patient into the research the medical researcher gives one of these envelopes to the patient as per procedure specified by protocol and without opening it (in case of double blind design) gives the studied drug corresponding to the envelope number. In case of urgent need (for example serious life threatening side effect) the research worker can open the relevant envelope and see what drug was used by the test person.

in case of accidental or deliberate opening of the randomization code the research worker must immediately notify this to the sponsor, specify the subject and the reasons of such decision/accident in writing. It should be noted that the advantage of the envelope method is relative independence of the research worker in the course of the patient’s randomization, which excludes dependence on the technical state of communication lines. However this independence is a disadvantage as well — the quantity of potential randomized test persons is limited by the quantity of available envelopes, the research worker must provide secure storage of the envelopes for prevention of unauthorized persons’ access to them, and the sponsor must regularly control the storage regime and the state of envelopes.

Recently centralized randomization by any method became widely spread. In this case the medical researcher notifies (on the telephone, fax or on the Internet) identification information and important data (generally the initials and age/year of birth refer to identification information; important data depends on the applied stratification — for example the main diagnosis or disease severity as per the relevant scale) about the test person to the department responsible for the randomization procedure for his/her allocation to a definite treatment group. The research worker receives a message in return concerning the treatment group or the number of the drug package (in case of double blindness of the research) for the ordinary patient, in order to ensure against errors verbal information is replicated by fax message or an e-mail. If disclosure of the code is necessary the research worker applies with the request to the department carrying out the randomization in accordance with the instructions received before the start of research. Completely automated telephone line systems – interactive voice response systems (Interactive Voice Response System — IVRS) or Internet network communication systems - interactive web response systems (Interactive Web Response System — IWRS) are often applied for carrying out the centralized randomization. The advantages of the centralized randomization methods are a higher level of protection from the unauthorized access, quick receipt of information about new patients into the research database, conventional unlimitedness of the quantity of potential randomized persons in the research centre, more ordered procedure of the code disclosure. The disadvantage is dependence on the technical state of communication devices and equipment.

When applying IVRS and IWRS systems every authorized member of the research team usually gets own unique code of access to the system. This code needs to be kept secret and must not be disclosed to colleagues. If your colleague of the research team left own envelope at home and asks you for help, you can carry out the randomization procedure yourself, but never tell your access code to him/her. This is important for correct maintenance of documentation and further reproduction of procedures if necessary (audit trail).


Q: What is "bioequivalence"?

The research of bioequivalence is a type of the clinical research with the aim to prove equivalence of two drugs – the original and its generic. For this purpose 18 (healthy) volunteers are engaged and after the drug intake their blood samples are collected in accordance with the expected curve of alteration of its concentration in the blood, the content of the generic and the originator (the compared drug) in the blood is determined and compared. On the grounds of these data the conclusion about the drugs’ equivalence is made.

Q: Are there any drugs which are not subject to registration?

Yes, there are, the following drugs are not subject to state registration:

1) pharmaceuticals manufactured by pharmacy organizations, veterinary pharmacy organizations, individual entrepreneurs, having license for pharmaceutical activity, as per prescriptions for pharmaceuticals and requirements of medical organizations, veterinary organizations;

2) medicinal plant raw materials;

3) pharmaceuticals purchased by individuals abroad of the Russian Federation and intended for personal use;

4) pharmaceuticals intended for export;

5) radiopharmaceutical pharmaceuticals manufactured directly in medical organizations in the procedure specified by the authorized federal executive authority.

Q: What are the types of randomization? What are their advantages and disadvantages?

There is fixed randomization (simple, block and stratified), dynamic allocation (method of “biased coin” and adaptive randomization). In case of fixed randomization the patient is allocated into this or that group on the grounds of random numbers, received from special tables or generated by a computer program. Simple randomization suggests equally probable allocation of test persons into the groups. So in case of two groups — basic and control ones, probability of getting into the treatment group is equal to probability of getting into the control group and is 50%. In this case sufficient difference in head count of groups, imbalance of groups as per age, sex, disease severity and other features can occur at a definite stage of research. Method of block randomization helps to achieve more balance between the groups as per the head count of test persons at each moment of research conduction — in this case randomization sequence is formed from blocks of fixed length, inside whereof random allocation is made.


Figure. Example of randomization sequence in case of block randomization.

Example of ready randomization sequence in case of block randomization of 16 test persons (fixed size of blocks) is shown at the Figure. «А» means allocation to A group, «B» — to B group, block size is 4, according to the protocol probability of allocation to this or that group is 50%. In this example the first randomized patient will be allocated to A group, the second and the third ones to B group and so on till the 16th patient who will be allocated to A group. The research worker has no access to randomization sequence and does not know to which group each next test person will be allocated.

However in case of block randomization the research worker can forecast to which group next test person will be allocated (if the block size, the previous allocations within the block are known, and one of two groups within the block is completely staffed) — for example it is obviously clear that patients 7 and 8 at the figure will be allocated to A group, if it is known that block length is 4, and patients 5 and 6 will be allocated to B group. In order to prevent such possibility one can use random determination of block sizes (with the use of random-number generator) or keep information about the block size secret if it is fixed.

Although the randomization principle, probability of allocation to this or that group, the applied technical method of procedure performance are described in the protocol of the clinical research, the protocol must not include specific details, which enable the research worker’s forecasting the result of randomization for a definite test person (for example the block length in case of block randomization). This requirement is contained in ICH E9 document.

In case of stratified (layered) randomization one or several (usually not more than two) important identifiers which can sufficiently influence the treatment results are taken into consideration, and therefore they must be evenly allocated between the groups. These identifiers can include sex, age, basic diagnosis, main pharmaceutical of basic (not researched) therapy, state severity at admission etc. This is made in order for specific samples (treatment groups) formed in this way to be representative in relation to general totality (of all test persons included into the clinical research) as per basic predictive factors, in other words, for every treatment group to be maximally similar to the general population of test persons of this research as per composition.

Method of “biased coin” enables achievement of more balance between the groups as per one definite identifier by means of dynamic change of probability of test persons’ inclusion into this or that group depending on the current group balance as per the fixed identifier. So the following algorithm will be applied for achievement of the current group balance as per the quantity of tests persons: in the course of test person’s inclusion into the research probability of his allocation to the group with the less number of participants will be more than 50% (as a rule probability of 66.6% is applied), and if the head count of the group at a definite stage is equal, probability of allocation into one of two groups for the next test person is 50%.

Methods of adaptive randomization are applied in case of adaptive design of clinical research, in which allocation of test persons into the groups is made so that more test persons could get the most effective (or the safest) drug or dose of the studied drug by the end of research.

In such cases probability of the patients’ allocation to this or that treatment group will dynamically change on the grounds of results of interim data analyses. There are many methods of response-adaptive randomization — for example “play the winner” method (Randomized-Play-the-Winner), utility offset model (Utility-Offset Model), maximum utility model (Maximum Utility Model).

The advantage of “play the winner” method is that more patients will get more efficient treatment. The disadvantages of this method include complexity of calculation of sampling size; necessity for outcomes of every previous test person to be determined before the moment of the next test person’s inclusion into the research; recurrent or constant data disclosure in case of blind clinical tests. Automation of the process of patients’ allocation to the groups by way of software development and gradual research conduction are applied in order to control these disadvantages.

In case of utility offset model application as method of adaptive randomization probability of the patient’s allocation to this or that group is calculated on the grounds of frequency of positive response for every treatment variant and results of the test persons already allocated to this group.

In case of adaptive randomization carried out by means of maximum utility model the next patient is always allocated to the group in which the highest efficiency of treatment is observed (or is supposed on the grounds of the model).

However there are specific difficulties and peculiarities of adaptive randomization methods application. For example in case of blind design it is necessary to provide recurrent or constant data disclosure (a separate group of “unblinded” statisticians is often engaged for this purpose); the speed of data analysis depends on the speed of their intake, that is why randomization of the next patient can take place before consideration of the previous test persons’ reactions etc.