The following text is a reproduction in most parts of the lecture I gave at the ceremony commemorating the twenty year anniversary of the foundation of the Controller Committee, held on July 24, 1992. Since then, many foreign experts have expressed interest in the specific details I talked about, and there were also requests for additional data. The lecture also provoked controversies among Japanese experts and lead to reviews of analysis of several thousands of points of data. Therefore, I have added some new information on the three following points to the text when the opportunity has carne to print out the manuscript of the lecture. Some say that such information is only pertinent in Japan, but I and the Committee as a whole would like to recommend that three points be adopted in practices in countries abroad.
The first point is the involvement of a third party in the field of evaluating efficacy, introduction of a double controller system, and the utilization of a computer system. Such a system may have no precedents in the world. However, it can be described as a type of social invention, incorporating ideas to allow mutual trust among various fields in the medical and pharmaceutical practices, in which conflicts of interests are inherent. The system has already been in operation in Japan for more than twenty years, and it has practically reached the realm of near-completion in terms of technicality and operation. The abilities of the system to maintain fairness in efficacy Studies are highly recognized and relied upon both in Japan and abroad.
The second point is the judgement of usefulness, which combines not only the separate aggregation of efficacy and safety information but furthermore combines effectiveness and side effect information of individual patients. A final judgement is made on each patient as to whether a drug was good, bad, reusable, or not reusable, and the results are comparatively Studied. Such procedures enable one to forecast how doctors and patients behave in prescription and treatment activities after the launch of a drug into the market.
The third point is the adoption in principle of non-parametric statistics as the method for statistical analysis of data. This is because the procedure is easy to comprehend for doctors, allows no undesired consequences that comes from manipulating data to meet the Standard level of information, and leaves less hypothesis in average, method of aggregation, and in statistical test method. Such characteristics lead to an advantage, that researchers both in the fields of medicine and pharmacology are less likely to be drawn into the qualm of statistical debate.
2. THE THIRD PARTY SYSTEM
In 1965, I became a member of the Subcommittee of New Drugs of the Central Pharmaceutical Affairs Council (of the Ministry of Health and Welfare), and attended an afternoon meeting once a week. At every meeting, the some old criticism was raised, that "only a scientific clinical report adopting a double blind trial method can be evaluated." Every time the secretariat replied that "the bureaucracy was in no position to give such guidance because there were few doctors with such experience, much less capabilities to give directions. Such a criticisms to the clinical practices were raised by members who had pharmaceutical backgrounds. Probably, they wanted to further enhance to the Standards of clinical medicine, and establish a methods which would produce efficacy data which would be trusted by non-medical specialists as well. However, there were precedents of a new drug evaluated by the double blind trial in Japan.
The double blind trial method was an innovative method, initiated in the United States around 1950 in the field of clinical medicine, to eliminate not only conscious biases but also some subconscious biases in pursuit of the truth, based on insights into individual and social psychology. This is a method which can tell us the truth, through the combination of random comparative trials and statistical testing in one form. When implementing this process, what is most important is the strict blinding procedure imposed by an independent party. The blinding procedures must be consistent through the subsequent data collecting and testing operations. If the system is made so that there would be no chance for later corrections or non-adoption of cases to intentionally affect the results in terms of procedures, then the process can gain the confidence not only of examining members and regulators who give approval, but also of the persons who are most important, the clinicians and all of the patients.
3. MAINTENANCE AND GUARANTEE OF THE DOUBLE BLIND TRIAL METHOD
BY THE CONTROLLER COMMITTEE (Figure 1)
Specific duties of the Controller Committee as a third party are described in detail in the flow chart(Figure 1). These include: a discussion of data from pre-clinical and preceding trial phases, designing, decision on the case card from, and considerations of the test drug, control drug, and their identifiability including packaging (indistinguishability). When these are completed, the boxes or bottles containing several hundred tablets for one person are brought into a room with a certain size at the Committee, and they are numbered at random for the subject patients. The controller confirms before the allocation of the drugs whether the tablets, capsules and packaging are prepared so that the test drugs and control drugs are not identifiable, extracts a box at random and sends it to a neutral agency for testing the quantity of each effective ingredient. The box is sent to participating hospitals, and the doctors allocate the boxes to their patients, beginning with younger numbers. In order to prevent any later corrections or re-supply, data from all cases are entered into the Committee computer, printed out for corrections of any typographical errors or irrational output, and if still there remain any doubts, then the central committee clinicians discuss these doubts. Then, the data is fixed after obtaining unanimous consent. At a meeting attended by company representatives and doctors, the controller breaks the blind and enters the key code in the computer. Then, following the program, all of the cases are separated into test drug groups and control drug groups all at once, aggregated and tested for any significant differences. The results are calculated during a meeting which lasts about two hours. The entire procedure and the outcome is open to all the participants according to "general meeting method," and furthermore, this method is quick and fair.
Although this method allows for rapid processing, it can be too thrilling for company representatives. Further, if the results show the inferiority of a new drug compared to the control group or the placebo, then, the following discussion meetings or subsequent dinners become so gloomily overcast by an air of depression that they resembles funerals. Therefore, this clear-cut, no-cheating method has since been discontinued, and has been replaced with a "two Stage method," in which a group of representative doctors or the central committee meet company representatives first to study the result printed out by the Committee and, after putting the information in order, they report the results at the general meeting. Through this method, all the participating clinicians are forced to trust the central committee and the company.
4. MAINTENANCE OF RELIABILITY THROUGH DOUBLE CONTROLLER SYSTEMS
For the past twenty years, the Controller Committee has adopted the double controller system, in order to instill confidence in the results of its data processing, while strengthening its independence from any conflicts of interest. In principle, a person from a company or a member who has been requested to do the processing work, and another person either from the Committee, or in some cases another person from the government or from a subcommittee of the Central Pharmaceutical Affairs Council, blind and codify the drug. Unless the key codes of these two persons are matched, the name of the drug cannot be identified.
We have decided to impose a multiple controller system, because even when there is a report claiming the superiority of a new drug based on tests are published, and even if they come with some added explanations, a specialist in basic medicine cannot simply agree to the result if there are problems of onset of action in terms of drug metabolism or pharmacology. This is because even if there is a claim of efficacy in the clinical practice, such explanations cannot be logically structured in terms of pharmacology and metabolism. In order to technically dissolve such a deadlock, we have decided to create a system so that each controller alone cannot break the code, or in other words, a system that does not enable one person to know the details of the drug allocated to each patient. This is the reason why we have decided to have another member of a subcommittee become a controller for re-trials. As a result, we now use the double controller system for the phase three trials. Also, it is possible for a company to blind the trial in advance before presenting it to the Committee.
The controller inspects the trial protocol as shown in Fig. 1, presents the protocol to the company which market the control drug and obtain their consent for a fair comparison. Furthermore, the controller is also responsible for random inspections in order to confirm the quantity of the contents and the distractibility of the test and control drugs, random allocation, distribution of drugs, data collection, communications, later tampering of data, aggregation and processing, selection of a test method, and making decisions on the output form. Further, when recordings of all of the scheduled clinical cases are completed (although a part of the following is omitted from Figure 1,), the controller may conduct a sampling test to check the Stability of remaining drugs. Finally, after publishing a report paper which includes the results of calculations following a consultation with the supplier of the control drug, the data is reconciled.
Obtaining informed consent from patients is a matter-of-fact criteria. Once in the past, this requirement was rejected because it was "impossible to explain placebo administration" to the patients, and the Committee decided to discontinue to cooperate with the project, in the middle of a process to prepare protocol.
5. STRUCTURE OF THE CONTROLLER COMMITTEE
The committee has been paying attention and making efforts to be responsible as a third party to ensure fair double blind trials, not only for the commissioned administrative work, but also to formulate methods which would produce reliable results of the comparative aggregations. Such complex and precise procedures as well as their checking and computer operations are not the type of work that an individual can perform alone, no matter how much a person may be skilled. d. If we intend to be serious about our work, then we must establish a third-party organization which would work according to the rules. These are (he reasons why the foundation of the Controller Committee as an organization has become necessary. Originally, a spontaneous gathering of interested doctors, it has grown to have its own secretariat and to employ full time workers who work on the neutral position with trained skills.
The members of the Committee mainly consist of clinicians of various departments, joined by lawyers as well as experts in the fields of statistics and epidemiology. Furthermore, implementation of decisions made by the committee is the responsibility of the powerful and competent ladies working for the secretariat. Their roles are not merely clerical duties to perform administrative works as directed, but rather they are expected to understand all of the decisions or the flow of work from the position of patients and their families; and, they are so trained in the beginning. Doctors, lawyers, statisticians and epidemiologists who are the members of the committee have grown to become authorities in their respective fields in the past ten to twenty years. But the ladies at the secretariat surely have become authorities as well, and their role as specific and specialized critics, not only of the industry and the government but also of the Committee members for their behavior or subtle attitudes have become too important to neglect.
In consideration of the strong demands on the clinicians from the governmental representatives with pharmaceutical background in the Japanese permit and approval agency as I have discussed at (he beginning, the Committee should invite participation by pharmacologists and pharmacists in the organization so that it becomes more socially recognized and Stabilized. Further, since trials of this kind have major impacts in the health of the public as a whole, it is necessary to create a system so that the government can be responsible for these activities. And when that is realized, or until then, it will be desirable that regulators, members of the Central Pharmaceutical Affairs Council, especially the members of the new drug subcommittees, accept the role at least as the second controller for confirmation and recognition of the fairness.
6. THE PUBLICATION: CLINICAL EVALUATION
At monthly Committee meetings, members can discuss the scientific rationality of a trial, its fairness and ethical permissibilities by each presented protocol, more freely and in detail than at congress meetings of clinicians. The issue of informed consent in experiments on a person has been discussed in the articles of Clinical Evaluation since early times. The magazine has been basically publishing all of the trials conducted by the Committee, and when questions have been raised concerning its fairness, follow-up articles were written.
Publications include results of trials which presented no differences from placebo, or cases when filing for approval to the Pharmaceutical Affairs Council was abandoned because of lack of any difference from placebo or of the inferiority than the control drug. In principle, all of the activities conducted by the Committee must be publicized, in order to maintain its reliability and to allow advancements in scientific methodologies.
What is worth paying attention to is that writing fees are paid for reports concerning side effects.
7. GENERAL IMPROVEMENT RATING; GIR
The Committee agrees with some economically advanced nations in such fundamental concepts as "the Logic" presented by the American IND, "Ethics" which carne out of Helsinki Declaration that was made in retrospect of the atrocities of the Second World War, and the "psychological" insights in the blind operations. The World believes that it has two unique characteristics in terms of specific handling of data. One is the importance it has attached to the final judgement of "usefulness," and the adoption of nonparametric statistics method which do not require normal distribution or ratio/interval scale as preconditions.
First of all, as seen in the study reports from many different countries, we aggregate and discuss various symptoms and changes in test values categorized into groups of new and control drugs. At the same time, we also summarize the degree of improvement or deterioration of symptoms by each patient in order to evaluate the degree of general improvements. The final general improvement ratings are made in multiple ranks as follows:
1. marked improvement
2. moderate improvement
3. mild improvement
4. no change
This rating is possible even when there is a lack of measurements or missing test values such as blood pressure during treatment. It is well known that when a condition of a person improves, it is not clear whether the improvement was a result of a drug's efficacy, the natural course of the disease, or is some psychological stimulus. w is certain is whether the condition has improved, remains unchanged, or unfortunately deteriorated. The factors of improvement cannot be identified when merely turning to individuals; only when the test drug and control drug groups are compared collectively, the factors of improvement brought by the test drug can be speculated in comparison with certain control drugs.
8. GLOBAL SAFETY RATING; GSR
However, a global safety level which relates to side effects and the occurrence of unexpected events is judged as well. In principle, side effects appear in areas that were normal before the treatment; in other words, side effects appear in the healthy area of the patients. Usually, incidences of side effects are limited and are distributed in multiple organs. The number of cases with rashes or headache are aggregated, and they are represented as a total number, and are compared between the test drug group and the control group. The Controller Committee considers that there are additional aspects to such data; when a rash appears on a face of a young female patient or when several disorders are recognized in a single patient, then the degree of somatic and psychological disorder for the patients would vary by individuals and are more complex. Therefore, a judgement should be made on the individual cases as to how the individual would accept the disorder. An expert who can give advise to a patient at such times would be the attending physician. There are patients who would simply discontinue administration without a word, or those who stop hospital visits. It is highly possible that such cases involve serious side effects. Discontinued hospital visits, without informing the clinicians, is itself important information. It could be because the patient has caused fire disasters, accidents, suicide attempts or crime, and there may be involvements of sleepiness, low blood sugar level, depression, blurring of consciousness or bad mood behind such incidents. Stopping hospital visits are included in the safety judgement by individuals, and the rating is made in three to five grade.
1. continue treatment with care
2. lower dose or add treatment of side effects, etc.
3. discontinuation, dropout, or death.
As mentioned in the section of general improvement. when a symptom which seems to be a side effect is recognized, often times it is not known in individual cases whether such symptom is caused by the drug or is an incidental outbreak unrelated to the treatment. In such cases, such symptoms should be recorded even when the incident is highly unlikely to be associated with the treatment as a side effect. Later, it may become a clear case of side effect after aggregation and comparison of data, or through post-marketing survey or pathochemical studies. Therefore, anything that was unexpected, whether it is a somatic symptom, disease, death caused by the disease, or psychosocial event, must be recorded. A patient's complaints of sleepiness may be judged as rate 2 requiring lower dosage if the patient is a professional driver or a student during examinations. But if the patient is a
retired person who can take a nap whenever he wants to, the rating may be 1 or treatment with care. An event may be handled as a neutral incidence when considering its relationship with a drug, but for patients who experience the event, it involves a value judgement. The social value of a drug varies from a country where the majority of citizens drive motor vehicles, to a country where automobiles are still rare possessions. The relationship between a drug and a patient who take the drug are comparable to the genotype/phenotype relations. This is the reason why the social value of DDT and BHC varies greatly from one country to another.
9. GLOBAL UTILITY RATING; GUR CLINICAL USEFULNESS
The global judgements of "general improvement" and "global safety" are the benefits and risks for individual patients. The degree of usefulness for each patient is judged at the end, taking these two factors into consideration. The attending physician summarizes the Smoothness and comfort in a patient's treatment as "usefulness", respecting the opinion of the patient. The words used to describe usefulness vary by the specialist or by efficacy project, but an example would be:
1. markedly useful
2. moderately useful
3. mildly useful
5. undesirable development.
Such a global judgement may be similar to rating of excellent, fair, and poor, which are often used to describe clinical developments: Therefore, this treatment evaluation is accustomed by not only doctors but also by patients and their family members as well. Here, the subjective view of a patient as a comprehensive being, or the empirical overall judgement by a physician, can be respected and aggregated. Further, such global judgements overlap with the principles of quality of life (QOL), a concept which has become widely known in recent years. A patient who dropped out and cannot be rated for his general improvement because only initial test data are available, can be rated by the usefulness. Even when the reason for dropout is unknown, the case could be rated as an "undesirable development." In other words, the judgement of usefulness can be made on all patients regardless of availability of information, and any serious abnormalities which occur during the course of treatment can be recorded in this category. In the case of a phase three trial, this comprehensive jidgement for each individual patient can be used to forecast the way the drug would be used after it is marketed. In general, changes in improvement, safety and usefulness ratings have not necessarily caused by the drugs, as mentioned earlier. They include natural changes or placebo effects. These are all included in the accurate recording and evaluation of the course and QOL by each patient. It is important to add descriptive explanations to the ratings as well.
For instance, in the case of a cholesterol lowering drug, if the total mortality rate is the same for the drug group and placebo group, then one can conclude that impact on life or death is the same for the drug and placebo, even when many of the causes of death are not to be considered side effects at the time of making judgements. As in the case of hypotensives, cholesterol lowering drugs may produce psychosomatic functions and cause depression. Some people may take drugs when they do not want a sudden death from myocardial infraction. If the dose is lowered so that blood cholesterol levels do not fall too greatly, then it might lead to a difference in the mortality rate. A lesson may be gained through a re-analysis of data.
The above summarizes the function of the global judgement rating in Japan. There are some criticisms that the system puts too much confidence in doctors. Such a criticism may be valid in an open comparative trial, but in the case of double blind trials where considerations are made to the full extent, the overall judgement by the physician and the patients would not be biased if somewhat scattered, and the result should be taken as reasonable.
10. NONPARAMETRIC STATISTICS
Another aspect to be noted is the adoption of nonparametric statistics as a principle by the Committee. Three decades ago, around 1969, criticisms were raised in Japan about making quantitative representation of questionnaire or behavior rating scales, insisting that human beings cannot be quantified. In a sense, this can be compared to the argument between believers of mathematics and literature. It is true that such literary works as "On the Origin of Species by Means of Natural Selection" by Darwin C.R., " Etymological Souvenirs" by Fabre J-H., and a number of works by Konrad Lorenz prove the superb capabilities of linguistical expression. Only audio and image presentations can excel words. The volume of information is tremendous. Information received by physicians who have dealt with patients who were cured or not cured by treatments are also tremendous. Holding negative feelings about simple quantification of such information is a natural psychological reaction, and rather indicates a sincere attitude.
Nonparametric statistical method is very suitable for holding discussions which arise from the conflicts mentioned above. Such logical classifications as nominal scale, ordinal scale, interval scale, and ratio scale by Stevens S.S. et al. can be quite useful in putting these arguments within an appropriate scope. Nonparametric statistics, which places its bases in the standards of various scales, are also called statistics of behavioral science, such as psychology, medicine and sociology. It does not matter whether there is a lack of continuity, absence of normal distribution, or an obscure zero point. Each scale has its corresponding Statistics and tests. In the normal population, a blood pressure or cholesterol value may present a normal distribution, but when a value higher than a certain level is subjected to hypertension or hypercholesterolemia, naturally distribution is no longer normal, making mean values or Standard deviations meaningless, and it can only be handled by ordinal scale. If there are specific demands, then mean value and other parametric data processing can be added. But adoption of nonparametric statistics as a basis gives an advantage, that from the initial stage of mathematical expression of data to mean value, aggregation and testing, there is no need to establish hypothesis, which could develop into points of controversies.
Furthermore, this method is relatively easy to explain and comprehensible to clinicians. I do not intend to introduce here such a grand argument that "easiness to understand is one of the requirement to establishing democracy", but it is an important goal to adopt a method which can be accepted by clinicians who Lend to repel statistics. Those who have the most detailed knowledge about the complicated clinical situations in respective fields, such as interrelations among various symptoms, order of onset, or changes in the severity, should participate in the discussion about the validity of statistics, its shortage and excess, utilization and developments. The logics of statistics lies not in the learning of statistics outside of medicine, but rather within all fields of learning and is even in logics one would encounter in every day life.
11. GENERAL PURPOSE PROGRAM FOR EFFICACY JUDGEMENTS
The committee created a general purpose program for efficacy judgements in 1972 and 1974 (1988 for multiplicity). In this program, input and output forms were standardized as much as possible. After the controllers bring in the key codes, the codes are entered into the system, and all the following process through the output two hours later is automatically programmed. Therefore, the burden on the Committee member who works as it controller is restricted only to 1) confirm that the drugs are not identifiable in formulation and packaging; 2) conduct random allocation; 3) keep the key code in Strict custody; and 4) attend the meeting for key breaking and enter the code into the system. Since the Committee is mainly organized of clinicians, there may be cases when a member must leave immediately due to changes in his patient's condition. In order to have the clinicians accept the grave responsibilities despite such demands, the time require of their attendance is limited to the minimum. Most of the rest of the work can be handled by the secretariat.
12. ACHIEVEMENTS IN TWENTY YEARS
The section of informational Materials were prepared by all the Staff in the secretariat in short notice.
Table 1 shows the twenty years of history of the Controller Committee and its organizational publication, "Clinical Evaluation." Drug related accidents and the activities by the government and the pharmaceutical industry are compared in chronological order to facilitate understanding.
Table 2 lists research activities, title of investigations, and the names of publications that carried the papers of reports conducted by the Controller Committee members. Furthermore, there is a list in chronological order of the total of twenty-four papers of researches and investigations carried out by academic congresses, universities, research groups, the Ministry of Education, the Ministry of Health and Welfare, and Science and Technology Agency for which the Committee has extended human, academic, technical or financial cooperation.
Table 3 is a table showing the number of trials, categorized by names of diseases or by efficacy. During the past twenty years, 746 trial projects (an annual average of 37) were processed. This indicates that several dozen trials have been held in almost all of the medical fields. The total of twenty-five recent projects of long-term control trials or post-marketing surveys are still on-going, increasing the total number of projects to 771.
Table 4 States the name of marketed drugs chosen as control drugs, the scope of the trial year and the number of subject cases for major diseases. For some trials, standard or control drugs have been changed several times. For cerebrovascular diseases, it states that seventeen trials use placebo as control, involving 2,016 patients in the trial. There are also tables for the following other diseases in this same order: hypertension, gastric/duodenal ulcer, schizophrenia, neurosis, depression, arrhythmia, osteoporosis and chronic hepatitis. In Japan, placebo trials are very limited for diseases which have distinct treatment drugs such as hypertension, gastric/duodenal ulcer, schizophrenia, depression and arrhythmia. However, placebo control are frequently seen in the trials for diseases with proven statistically significant differences though with less difference in efficacy, such as cerebrovascular diseases, neurosis, osteoporosis and chronic hepatitis. However, they are more often a three-group comparison among the test drug, Standard drug and placebo rather than only a two-group comparison between the new drug and placebo. Some numbers do not correspond with those in Table 3; this is because some limited number of control drugs have been omitted, and there are some three-drug comparisons within the same projects.
Figure 2 is a chart showing the superiority/ inferiority of a test drug and control drug. In the overall trials, test drugs were superior with statistically significant difference in general improvement rating in 34% of the cases; They were inferior in 4% of the cases with statistical difference. In comparison with placebo, 529S of test drugs were statistically superior. However, in 46%, nearly a half of the cases, there were no significant differences. Furthermore, the test drugs were inferior in 29& of the cases, including coincidence. In safety, test drugs were significantly safer only in 13%, while test drugs were inferior in 12%. Needless to say, active test drugs were inferior to placebo in safety in 22%. However, superiority in safety to placebo was seen in 4%. Usefulness results are almost identical to that of general improvement ratio, unless there were serious safety problems. However, when placebos were used as control, side effects, etc., were taken into consideration and while the general improvement rating was 52%, usefulness was a lower 48%.
Table 5 is the superiority/inferiority chart (general improvement rating) by disease or by efficacy. There are six less trials than the 502 trials presented in Material 6. This section indicates the superiority/inferiority of the control drug (active) by disease or efficacy. For antidepressants, there has been no drugs excelling the standard drug for the past twenty years.
Table 6 is a superiority/inferiority chart with placebo control in general improvement rating, also by disease or efficacy. These are the details of trials presented in Material 6. Negative data from these trials have been publicized in the "Clinical Evaluation."
13. THE PAST TWENTY YEARS OF THE CONTROLLER COMMITTEE
When we first published our publication, "Clinical Evaluation coordinating Doctor Naotaka Shimizu thought that the magazine would be discontinued after three issues. However, it has lasted for twenty years, and it is necessary to think of the reasons why. Such consideration would also highlight the reasons for existence and continuation of the Committee and its system from a different perspective.
First of all, there were very few comparative trials, much less double blind trials, carried out in Japan twenty or thirty years ago. The term double blind trial was used (by Greiner) in the united States for angina in 1950, forty years ago. It was in 1950 and 1951 when reports were made of objective trials with allocation by third parties, described as carefully controlled trials, for a trial which denied the intelligent improvement effect by glutamic acid. As a result, in Japan as well, the so-called administrative guidance was made and since 1965, the new drug subcommittee has become the main body to bring up the level of clinical evaluation, and it has given advice to implement the trials to companies which have shown interests.
Secondly, many pharmaceutical companies, with an understanding of the strict Committee system as an implementing organization, have respected the evaluation results by an independent party. Though it may not be the industry as a whole, there are many people who hope that the company they work for do not indulge in disgraceful acts such as manipulation of data. whether consciously or subconsciously.
The third reason why the Committee should continue is the continued support from clinicians, if not all of them. Between 1970 and 1980, experiments on people have become a controversial issue at universities or congresses, and criticisms were raised about their wrong relations with pharmaceutical companies; however, these criticisms were only whispered among doctors and were rarely known to the public. Now doctors are requested to speak about evaluation of clinical trials; an act which in the past had been left up to company representatives. Papers and translated materials carried in "Clinical Evaluation" are used as reference materials for these evaluations. When the Committee is involved as an independent. third party to stand in between researchers and companies. it leaves less room for the doctors to become a target of criticism by medical students as "pawns" of companies. Furthermore. in the standards of measurements. statistical arguments, rationality of design, or reliability of processed data. the c can Committee a neutral position.
The fourth reason is that as a background of all these circumstances there are situations in Japan that the Pharmaceutical Affairs Bureau (of the Ministry) employs very few doctors. (only the director of the Biologic and Antibiotics Division has a doctor's title), and the manpower and budget are too limited. Moreover, many of the Japanese pharmaceutical companies employ only a limited number of clinicians. Therefore. doctors at university medical departments have little interest in drug developments or clinical trials. Also, clinicians are too busy to treat patients under the medical insurance system, and there are not very frequent innovative drug developments. These facts compounded by some past experiences, have created an air that doctors prefer to keep distance from companies and drug development activities. Also. voluntary reporting of side effects by clinicians is rather rare in Japan. All these may be a circumstances unique to Japan.
14. FUTURE PROSPECTS
Despite these circumstances, or perhaps because of these circumstances. the Committee can cooperate in creating an evaluation system for mid to lower stream that can also be a contribution to the up-stream, state-of-the-art research activities. for the committee to be active in the future and obtain support from a number of fields.
In the future. we hope to create a good database and filing system which would put the vast amount of information in good order, so that we can provide a regular database service which is also accessible to the general public. Then, it will be necessary to review various data by disease and by efficacy. Any clinical questions which may arise from such reviews must be addressed through discussions involving experts of the concerned diseases, pharmacology and pathology, to identify healing processes, mechanisms, or cases of improvements and incidences of side effects. we have accumulated a large quantity of quality data based on neutral and fair case reports, which we can be proud of to the world. There is a large volume of e for us to reflect upon, latent structure of diseases, and hints for future improvements and developments in the database.
15. FOR STRUCTURING RELIABLE SOCIAL SYSTEM AND ITS DEVELOPMENTS
In 1983, after I resigned as a member of the new drug subcommittee, I have emphasized the necessity to create an independent system at a lecture meeting of the Society of Japanese Pharmacopoeia. I stressed, that considering the situations in Japan where efficacy reports were made only either by pharmaceutical manufacturers or upon their requests, it was necessary to create a system to allow an independent evaluation, which can gain greater confidence of doctors, pharmacists and the public. Finally this year, a note about the role of the Controller was stated in the "general guideline," noting that "it is also desirable to assign controllers to the works which require neutrality." This is a Step forward. In the future, the specific meaning of this word would become a subject for discussion, but unlike statistical testing, this argument is something which is easy to understand even if one is not a specialist. Therefore, the discussion must be open to the public.
The responsibility to implement fair and reliable efficacy evaluations lies primarily in the hands of pharmaceutical manufacturers. However, some unethical deeds are sometimes reported in the media. Therefore, some type of voluntary and private mutual regulation must be installed by the industry organization. At the some time, the government is also responsible to establish a system under which companies are required to make pertinent evaluations, or a system of self-evaluation. However, doctors or doctors' association and congresses that are the users of the products are also capable of making independent and voluntary efficacy evaluation if they wish. And if the results of such evaluations cannot be trusted, or even if they are trusted, patients, their families and the general public who are either taking the drug or who have the possibility of taking the drug in the future may decide to initiate their own evaluations.
Different people can evaluate drugs from different standpoints, and the Committee has operated as an independent organization for the past twenty years while giving considerations to the requirements of industry, administration, doctors, lawyers and statisticians, with a possibility of further adding new fields in the future. This methods can be quite responsive compared to the so-called third-sector methods, with a very low need for a new investment. Although the final authority lies in the hands of administrators, the type of data desired by the Central Pharmaceutical Affairs Council or a new drug subcommittee reflect their requirements when the controller is involved, and the fairness can be confirmed. However, there still remain some doubts and controversies, which are described hereafter.
First of all, there is a claim, that evaluation does not necessarily have to take place very strictly in the first stage, because drugs are subject to re-evaluation and possible elimination in phase-four studies or through post-marketing surveys. Drugs, among many other merchandise, are directly associated with human life and well-being, and the cost of a drug is covered by the insurance funded by the contribution and tax money paid by the public. It would be inhumane to continue a large-scale, uncontrolled trial without limitations to introduce a drug which had only received skeptical evaluation. Therefore, it is necessary to conduct investigations in the beginning as thoroughly as possible.
Secondly, in Japan, additional investigation of efficacy is difficult. Due to efforts to come up with Statistically significant differences even when the difference in efficacy rate is small. several hundred patients are involved in one double blind trial. In such a case, additional investigations after marketing would not be a simple task. In Western nations, "the blinding method" is adopted for anti-myocardial infraction drugs or anti-infraction drugs for a duration much longer than the trial period by the time of approval. In Japan, however. there has been no precedent of additional investigations of similar scales. When considering the economy of the works and the inability to re-try the clinical experiment, it is vital that the initial comparative trial bears great importance.
Thirdly. the market is structured in Japan so that the mechanism of market principle does not work even after marketing. It may be too extreme to say that there is a neglect of usefulness for the patients: nevertheless, the system is made so that the difference between the national health insurance prices set by the government and the price at which dregs are actually sold by pharmaceutical companies is a decisive factor of superiority/inferiority and the share of the drug in the market. Further. because it is the piece work payment system, users have a tendency to do more clinical testing and to prescribe multiple drugs which have larger price differences. In this case. drugs with low effect comparative to that of placebo are more welcomed and adopted in multi-drug prescriptions. Doctors shy away from drugs with quick action with sufficient potency with a single drug. Theoretically speaking, the Japanese health insurance coverage system involves a danger of creating such a tendency. Still, however, the Japanese national health expenses are contained. supporting long life expectancy and long working hours, because of very close regulations installed by the Health Insurance Bureau, and the expertise and the sound sense of clinicians as specialists.
The fourth point is that as stated in the subconscious bias or in the defence mechanism by Freud. S. a double blind trial method is basically premised on the involvement of an independent organization. It may be too extreme to speculate that a company which has developed a drug and is an interested party may designate a controller, and handle all the time consuming minor works for the controller as if the company is only borrowing the name of the controller. Still, the procedures should be made more transparent. Also, the time consuming minor works to be handled by doctors or the doctor responsible for the overall trial activities, may be something that they would like to stay away from, but in reality, these cumbersome miscellaneous works have significant importance because they are subtle and are likely to affect the results of evaluation. It would like putting a fifth wheel on a coach without any assurance of fairness, when only giving detailed procedural restrictions through administrative guidance, and merely checking the system additionally.
The above text is the summary of my own assertions and the members of the controller committee concerning the methods of fair evaluation of Japanese and foreign drugs, expressed at the commemoration of the twenty year anniversary of the foundation of the Committee.