Evaluation research can explore whether Tibetan physicians, using traditional methods of diagnosis and treatment, are successful in treating various health problems, or in helping to prevent them. However, designing studies that will truly and fairly address that question presents a real challenge. Specifically, studies that impose restrictions on various aspects of the Tibetan doctor's work are inadequate for assessing the full potential of Tibetan medicine.
Our goal here is to design a research project that could show whether the Tibetan system of medicine -- as defined by the preferred diagnostic and therapeutic practices of respected Tibetan physicians -- is truly effective, using Western standards of evaluation.
To fairly test the efficacy of Tibetan medicine, we must make sure that the treatments we are evaluating are the best that Tibetan medicine has to offer. Obviously that will not be what we are studying if our research design interferes in the relationship between the Tibetan doctors and their patients.
In the research desigh we are proposing here, Tibetan physicians would be entirely in charge of how their diagnosis and treatment is done. Western scientists would be in charge of the evaluation statistics, and Western doctors would provide medical diagnosis before and after the treatment of each patient, so the results will make sense from the perspective of Western scientific medicine. Including Western diagnosis will also allow us to begin to study traditional Tibetan medical diagnosis, by comparing Tibetan diagnosis and Western diagnosis of the same patients.
A commitment to respecting the Tibetan doctors' standards of practice will challenge our ingenuity. If the Tibetan physicians are to be entirely in charge of how their diagnosis and treatment is done, we obviously cannot require them to merely pretend to treat half of their patients. Thus we will have to set aside the experimental plan often used for clinical evaluation studies: We cannot use a randomized double-blind placebo-controlled clinical trial.
By reviewing the reasons for the various aspects of the standard research protocol we were able to come up with an alternative design -- one that would preserve the validity of the study without interfering in the relationship between the Tibetan doctors and their patients.
In a standard clinical trial the control condition makes it possible to determine whether it is the treatment we are studying that caused whatever improvement we may see. If we are to eliminate the control condition, we will need another way to determine that it was the Tibetan treatment and not the mere passage of time, or some other factor, that caused the improvement.
What we propose, as an alternative to the standard double-blind placebo-controlled experimental clinical trial, is an observational study, comparing the results of Tibetan treatment with expectations based on medical statistics. For each of the diseases included in the study, medical statistics can be used to predict the liklihood of various outcomes -- complete recovery, recovery with disibility, or death, for example -- if the patient is not treated, and also what the outcomes are likely to be with various forms of treatment. By comparing the results of Tibetan treatment with these predictions, we can determine if Tibetan treatment for a particular disease is as effectiive or more effective than other treatments, or if it is no more effective that no treatment at all. We can also compare rates of negative side effects, costs, and other relevant variables.
top of page
When we in the West hear that Tibetan medicine can cure diseases like cancer and mental illness, and do so without lasting harmful side effects, we are naturally skeptical. We want to see the evidence that these treatments actually work.
In Western medicine we have a long tradition of clinical research devoted to testing the effectiveness of various medicines and other forms of treatment. Many treatments that were once considered effective have been shown to provide little benefit, or to be actually harmful, when evaluated scientifically.
Evaluation research can explore whether Tibetan physicians, using traditional methods of diagnosis and treatment, are successfull in treating various health problems, or in helping to prevent them. However, designing studies that will truly and fairly address that question presents a real challenge.
In an editorial titled Tibetan Medicine in Prime Time, I discussed some of the research that has been going on recently, studies aimed at evaluating particular Tibetan herbal formulae as medications, or merely as sources for new drugs, with no involvement of Tibetan physicians in diagnosis nor in prescribing medicines. I argued that if our goal is to evaluate the real potential of the Tibetan medical system, studies of that sort are inadequate.
Furthermore, even research that does involve Tibetan doctors may fail to fairly assess the full potential of the traditional Tibetan medical system, unless the Tibetan physicians are allowed to practice medicine in whatever ways they consider most appropriate. Many studies have imposed restrictions on various aspects of the Tibetan doctor's work.
Our goal here is to design a research project that could truly show whether the Tibetan system of medicine, as defined by the preferred diagnostic and therapeutic practices of respected Tibetan physicians, is truly effective, using Western standards of evaluation.
To fairly assess, scientifically, the full potential of Tibetan medicine, we must make sure that the treatments we are evaluating are the best available -- and for this we must rely on Tibetan doctors. However, we can assume that the treatments they are using are the best that Tibetan medicine has to offer only if our research design does not interfere in the relationship between the Tibetan doctors and their patients.
Tibetan doctors naturally strive to give their patients the best possible treatment, so any restraints we place on the Tibetan doctors choice of medicines, or any other interference we impose in the relationship between the doctors and their patients, will mean that the treatment we are evaluating is not the best that Tibetan medicine has to offer.
In this type of research the Tibetan physicians would have to be entirely in charge of how their diagnosis and treatment is done. Western scientists would be in charge of the evaluation statistics, and Western doctors would provide medical diagnosis before and after the treatment of each patient, so the results will make sense to us. Including Western diagnosis will also allow us to begin to study traditional Tibetan medical diagnosis, by comparing Tibetan diagnosis and Western diagnosis of the same patients.
The overall outline of such a study is simple; however, working out the details of how to make such a study truly informative presents some real challenges.
The first challenge we face comes from our presumption that we already basically understand what is important in Tibetan medicine and what can be safely discarded. If a traditional medical system, combining spiritual, psychological and material practices, actually works to improve some serious health problem, we easily assume that it must have been the material elements of the system, the medicines or physical manipulations of the body, or perhaps the physiological consequences of changes in behavior or diet, that caused the improvement. Accordingly, with passionate faith in scientific materialism, we further assume that there is no need to include spiritual aspects of the system in a test of it's efficacy, since they can't really be making any difference.
This does presume too much, however, as I think you can see. If we are trying to enter into a dialog with the Tibetan physicians, then it is not enough to translate their terms into our language. We need to add Western diagnosis to the study, in order for it to make sense to us as evaluating treatments for diseases that we recognize; but we also need to include whatever the Tibetan doctors want to include in order for them to accept it as a scientific study of the full potential of Tibetan medicine. For example, if the doctor wants to prescribe mantras to the Medicine Buddha, and the patient wants to say them, who are we to insist that they must be irrelevant? We can study compliance in saying the mantras in the same way that we study compliance in taking the medicines, without much extra effort.
If we do find that Tibetan medicine is effective for certain types of illnesses, then interested people could certainly perform follow-up studies to see if it might be just as effective without the mantras (or without the herbal medicines). At first, though, we can avoid possible confusion about whether we have really done a fair test of the Tibetan system of healing, by respecting the Tibetan doctors as experts on issues of how Tibetan medicine is practiced.
A commitment to respecting the Tibetan doctors' standards of practice will challenge our ingenuity. For example, if our research design is to impose no restrictions on the doctors' choices of herbal medicines, that might require waiting to do the study until the FDA has approved all the tibetan medicinal formulas that our doctors would like to be able to use; or it might mean that the study would have to be done in India or some other country that allows Tibetan physicians the full use of their traditional pharmacopoeia.
We also will need to develop a way to provide Western diagnostic services for the patients, including interviews and lab work, under the direction of Western physicians. These services will be needed before treatment, to establish the type, severity and complications of the patients problem (from the Western perspective), and after treatment, to determine whatever improvement may have occurred.
We can deal with these challenges; not easily, perhaps, but straightforwardly. However, one major challenge remains: If the Tibetan physicians are to be entirely in charge of how their diagnosis and treatment is done, we obviously cannot require them to merely pretend to treat half of their patients. Thus we will have to set aside the generally accepted standards for clinical evaluation studies: Specifically, we cannot use a randomized double-blind placebo-controlled clinical trial.
We will need to review the reasons for the various aspects of the standard research protocol, to see if we can come up with an alternative design -- one that would preserve the validity of the study without interfering in the relationship between the Tibetan doctors and their patients.
In the next section we will discuss the detailed logic of a research plan that seems to meet both those criteria.
top of page
Specifically, for this purpose randomized double-blind clinical trials are overly expensive, overly focused, and overly disruptive of normal clinical practice, when compared a less disruptive design that we will discuss in some detail. However, before we would want to seriously consider adopting such a design, whatever its virtues, we would need to be assured that it would provide valid evidence of the effectiveness, or lack of effectiveness, of the medical system we intend to study. So we will need to consider some of the specific characteristics of the standard clinical trial, to see what criteria any alternative design must meet.
In a typical randomized double-blind clinical trial, patients who present with a particular condition, e.g., people with the symptoms of diabetes mellitus, are randomly assigned to either of two groups. Those assigned to the "treatment" or "experimental" group are given the medicine or other treatment that is being evaluated, and those assigned to the "control" or "comparison" group, are given a placebo treatment -- a sham treatment administered in the same way, by the same personnel, as the test treatment, but having no physiological effect. If possible the study is double blinded, so that neither the physician nor the patient knows what treatment the patient is receiving, although both are aware of other aspects of the study. In research governed by modern Western ethical standards for medical research, the patient and the physician must both understand that they are participating in a clinical trial that may randomly deny treatment to the patient.
In randomized double-blind clinical trials, the procedure for obtaining informed consent for participation in the study is just the first of many disruptions in the normal operation of the clinic and the normal relationship between patient and doctor. Uncertainty about whether the patient is receiving real treatment or a placebo can undermine the atmosphere of healing generated by the environment of the clinic and the psychological and spiritual bond between the patient and the physician. The more serious the illness, the more this aspect of the study may tend to make the patient anxious, and undermine the doctor's confidence that she is doing her best to help the patient. Taken together, we might think of these influences as a Negative Placebo Effect, produced by the requirements of the randomized double-blind research plan.
The design we are proposing eliminates the control condition. All patients will be given whatever treatment the physician considers appropriate. There are no placebo treatments and no procedures for withholding information from patients or physicians. No physician is asked to pretend to treat some of her patients. With each patient she does the best she can to provide appropriate treatment, and both she and the patient understand that.
If we want to eliminate the control condition from the design of a clinical trial, we need to consider why the randomized double-blind design has become the standard for research of this type, and whether we can find other ways of accomplishing the objectives that led to its adoption.
Randomized: Random assignment of patients to treatment and non-treatment conditions.
Controled experiments are so common in scientific work in almost all fields that many people erroniously believe that all scientific research involves them. Actually, non-experimental observation is an equally valuable method, which is especially useful in beginning to explore a subject. (In one of my essays on science and its methods, titled Dispelling Some Common Myths about Science, I discuss this topic in some detail.) What I am suggesting here is that an observational study would be more appropriate than the standard experimental clinical trial in the initial phases of studying Tibetan medicine.
In a standard clinical trial the control condition makes it possible to determine whether it is the treatment we are studying that caused whatever improvement we may see. People's medical problems can improve for all sorts of reasons -- a change in the season, a change in diet or behavior, or simply the passing of time, which works quite well in resolving the common cold and many other medical problems. If we randomly assign patients with a particular diagnosis to either a group that will be given the treatment we are studying, or a group that will not be given the treatment, we can compare the two groups to see if the patients who were given the treatment improve more than the untreated patients.
In the study we are proposing, we do not need a control condition to determine whether patients have improved. Assessing whether any given patient has improved during the period of the study can be accomplished by comparing the results of the standard Western diagnostic measures that will be taken before and after treatment.
However, in our study, just as in other clinical trials, it will take more than improvement in a patient's condition to demonstrate that the medicine we are studying has helped them. Including a non-treatment control condition would have allowed us to determine whether it is the Tibetan doctor's treatment that causes whatever improvement we may see. However, since we want to avoid the control condition and the Negative Placebo Effect it may cause, we have to come up with some other way of determining whether Tibetan medical treatment is contributing to whatever improvement we may see in the patients.
There is a way to do that: We can determine statistically whether the Tibetan medical treatments were an effective curative agent by comparing the results from patients in the study with historical evidence of what would be expected in people with that particular diagnosis. What percentage of people with that illness would recover, what percentage would continue to suffer the condition chronically, and for what percentage the problem would be fatal or have other undesirable consequences -- we would need to find such statistics in the medical records for various regions. The records may show what outcomes can be expected both without treatment and with standard Western medical intervention.
This is not a new idea. It is what is normally done when placebo- controled studies are considered ethically inappropriate, for example in research on life-threatening diseases where a somewhat effective treatment is already available, and researchers are testing alternatives that might be more effective.
In preparation for a study evaluating the Tibetan medical system, Western medical consultants would have to determine, for each type of health problem, how much improvement on the diagnostic indicators would be required to demonstrate that the Tibetan treatment had indeed produced a therapeutic effect. For example, Western physicians might consider it quite unlikely for a particular type of tumor to disappear from a patient's x-ray without effective treatment. So if a group of patients are given Tibetan medical treatment, and in a large percentage of them the tumor shrinks or vanishes, we can safely conclude that the treatment caused the improvement in most of them, since only a very small percentage of these tumors would normally disappear spontaneously. In contrast, for an illness that will normally clear up without treatment, we might have to see an improvement that was clearly faster than normal, or which left fewer lasting negative effects such as loss of function or disfigurement, before concluding that the treatment had a real therapeutic effect.
The main disadvantage of the base-rate comparison we are proposing is that it is less sensitive than the standard randomized treatment / non-treatment comparison. However, the proposed design offers a way to make a very cost effective preliminary evaluation of the traditional Tibetan treatments for a wide range of medical problems. Any results that are dramatic enough to inspire further study can be followed up with more sensitive experimental studies.
Placebo (Sham Treatment) Control
Patients assigned to the control condition in clinical trials used to be simply given no treatment, but this is no longer generally considered appropriate. To control for the effects of a phenomenon known as the "placebo effect" patients in the control group are given fake treatment which is as similar as possible to the real treatment except for omission of the therapeutic agent being studied: sugar pills, injections of slightly salty water, whatever will most resemble the actual trealment without having any physiological effect is given to the patients in the placebo control condition in the same way that the active medicine is administered to the patients who are receiving the real treatment.
"Placebo Effect" is the term commonly used for improvement that results from the psychological effects of believing that one is being given an effective treatment. The Placebo Effect was believed to result in measured improvement on various physiological indicators, and it even makes sense in light of evidence showing that improvements in a person's mood can have a positive effect on the immune system. However, some people have raised the question of whether a placebo effect exists at all.
Before deciding to use a placebo-controlled randomized experimental design in a clinical trial we should be clear about exactly what we intend to study. If we are focusing solely on the therapeutic value of a particular medication or other treatment, then radomly assigning half of the patients in the study to a placebo control condition may be a straightforward way to isolate the effects of the treatment from other known and unknown factors that could lead to improvement in a patient's condition. However, if what we intend to study is an entire medical system, one that we don't understand, then we have to study the system as is is traditionally practiced, at least at first.
Specifically, if the system we intend to study is traditional Tibetan medicine, in which psychological and spiritual influences are intentionally included for their synergistic enhancement of the therapeutic effects of medications and other treatments, then the Placebo effect, or something very much like it, is part of the therapeutic process, and is not in any sense something we should be trying to control for. In an initial study of the Tibetan medical system, subjecting all the participants in the study to the Negative Placebo Effect we discussed earlier is clearly inappropriate.
If a preliminary survey of the effectiveness of the Tibetan medical system, in treating each of the different medical conditions that it is commonly used for, finds that it does actually offer some benefit for patients afflicted with one or more of these conditions, we will of course want to know how it works. For each particular medical problem, we can ask which of the many elements of traditional Tibetan medical practice actually make some positive contribution to the therapeutic effect demonstrated in the preliminary studies.
In research aimed at answering those questions, placebo controls may have some role to play. However, the design used in standard clinical trials will not allow us to answer the question of whether something like a placebo effect mught be synergistically enhancing the effects of other aspects of Tibetan medical treatment. Do do that would require a design with four different groups: Treatment, No Treatment, Placebo, and a group that is given the treatment withiout being aware that they are participating in a randomized clinical trial in which some of the other patients will be given no treatment or placebos. This design would also allow us to determine whether the standard design does in fact produce what I have called here a 'Negative Placebo Effect."
Another issue is the absence of double blinding in our study. Since there is no untreated control group there is no need for keeping the patients and the doctors ignorant about what is going on. However, double blinding also avoids any bias on the part of the physician in evaluating the results of treatment. In our pre-test / post-test design such biases are prevented from affecting the results by the fact that the Tibetan physician is not involved in the Western diagnostic assessments, except in one way -- deciding when the patient has been cured or has improved sufficiently so that the post test diagnosis should be given.
In addition, it would be fairly straightforward to make sure the technicians in the lab or x-ray facility are not even aware that the procedures are part of a research project. However, the possibility of bias influencing results is present whenever people perform diagnostic procedures involving contact with the patient and direct assessment of relevant variables, for example temperature and blood pressure readings, vision tests, neuropsychological assessments and any of a host of other procedures. These people, along with physicians making overall assessments of patients clinical status, could be made aware that the work they were doing was part of a research project and that their findings would not be involved in the treatment of the patients, but they and those running the statistical tests could remain unaware of the nature of the study until it is completed.
In later follow-up studies which may appropriately involve control groups, it may be impossible to blind the subjects regarding certain aspects of the treatment such as dietary and behavioral changes, performance of prayers to the Medicine Buddha, exercises, etc. Furthermore, although it will be possible, in general, to ensure that the people evaluating the outcome from the perspetive of Westarn medicine are unaware of the treatment, it will be difficult if not impossible to blind the Tibetan physician as to whether the patient is recieving real medicine or placebos. Traditional practice includes re-examining the patient after a few weeks to assess the effect of the medicine and adjust the prescription if necessary -- and to reexamine the patient and consider readjusting the prescription whenever there is a notable change in symptoms. Tibetan physicians believe that they can tell by pulse analysis what effect the medicine is having, and that would of course include telling if the medicine is having no effect at all.
Comparing the results from patients in the study with historical evidence of what would be expected when people with a particular diagnosis are not treated, and when they receive standard Western medical treatment, will allow a survey of the relative effectiveness of Tibetan medicine for a wide range of different medical problems. Tibetan medicine might come to be considered the prefered treatment for a particular condition because it is more effective than the available Western treatments, or because it is less expensive than equally effective Western treatments, or because it is less likely to result in serious negative side effects. All of these possible outcomes can be considered for each type of medical problem that is included in the study. However, even though this design would be much more cost effective than standard clinical trials, the expense of the Western diagnostics will require some care in selecting the specific maladies to be included in the study.
We can easily gain a rough estimate of the types of medical problems where Tibetan medicine is most likely to make a real contribution to modern medical practice by looking at what is happening in the few places where both Western and Tibetan medicine are available, places like the town of Dharamsala in India. In general, Western medicine tends to be the treatment of choice for acute illness and emergency problems, whereas Tibetan medicine is preferred for longer-term and chronic problems -- but what we need is a list of the specific conditions for which Tibetan medicine is considered the treatment of choice when both systems are available.
The first stage of this project should be to interview Western and Tibetan physicians who have been collaborating with each other in these places, to develop a list of diseases and other health problems that are most promising for inclusion in this study. Another important source would be Jürgen C. Aschoff's Tibetan Medicine - Tibetische Medizin, an annotated bibliography of books and articles published in all languages except Tibetan from the year 1789 up to 1995, including an extensive collection of reports of Western medical research.
After selecting a list of medical conditions to be studied, historical data can be consulted for the particular regions that will be included in the research. For each disease or other problem which is found to be included in the appropriate databases, statistical consultants can determine what size the sample of patients would have to be to show a statistically significant improvement over baserate outcomes, given the estimated average effectiveness of the Tibetan treatment for that particular poblem.
Potential Sampling Problems
Since patients will not be randomly selected for treatment, the research team must take care to minimize biases in sample of patients that are included in the study. For example, in areas such as Europe and America, patients who come to a Tibetan clinic are certainly not representative of the population as a whole, but are generally more affluent and better educated, and have better nutritional resources and better access to sophisticated Western medical care. For this reason, it might be better to begin the research in an area where Tibetan medicine is available to the entire population -- like some of the Tibetan refugee settlements in India.
Even in such a community, sampling biases can easily arise. For example, if patients are expected to pay for their own medicine, as is often the case in these settlements, relatively affluent people are more likely to appear at the clinic as patients. For this reason, it might be best if the funding for the research project could include resources for providing medicine free of charge to all the patients in the study. If that is not possible, at the very least such biasing factors as are likely to have influenced the inclusion of patients in the study should be carefully reported.
top of page
"Now, the Tibetan doctor takes not only the pulse, but the warmth, the
degree of tautness in the vein the quality therefore of the blood, thick
or thin, not only the rapidity, but other more subtle vibrations that are
transmitted by the pulse from the interior of the being and then is reaching
in with the cultivated ability to visualize into the liver, down the channels
of the biliary duct, and the gallbladder, into the stomach, into the throat,
into the heart, into the lungs, into the kidneys -- and is reaching by
different sensors, a set of 12 sensors, that are in the 2 sets of 3 fingers
that they press the pulse -- are different depths, and like the safe cracker,
they can hear something like a little flaw in an artery -- too much plaque,
and stiffness and rigidity in the way the blood spurts from the heart pump.
They can hear these internal things to an extraordinary degree. And they
build that up, by not only their own personal knowledge of the anatomy
-- but also it comes in a long empirical tradition that's come from thousands
of years of people observing different types of pulses, and correlating
those observations to different types of internal phenomena, symptoms,
and so forth, in their patients." -- Robert Thurman, Professor of Indo-Tibetan
Studies, Columbia University.
of the Planet, September 2001
Participants in the discussion of "Research Methodology in Tibetan Medicine" at the First International International Conference on Tibetan Medicine (Washington, D.C., November, 1998) expressed interest in studies of Tibetan medical diagnosis, as distinct from evaluations of the efficacy of Tibetan medical treatment. The design proposed here offers an opportunity to begin studying Tibetan clinical diagnostics, by comparing the diagnostic judgments of the Tibetan physician, both before and after treatment, with Western diagnosis, for each patient included in the study.
Such a comparitive study might give us a basis for confidence that when we talk with Tibetan physicians about diabetes or some other problem we are talking about the same thing. However, finding that Tibetan diagnostic classifications differ from Western classifications would be interesting as well, if the evaluation research shows Tibetan treatments to be effective.
In terms of the goals stated earlier in this proposal, of designing a study that would not interfer in the relationship between the Tibetan doctors and their patients, requesting that the doctors share their diagnostic conclusions with us would seem to pose no real problem. Although trying to discuss their diagnosis with us might tend to bend the Tibetan doctor's categorizations to correspond more closely to the Western classification system, that is a process that is happening anyway in as Tibetan physicians work with Western patients and colaborate with Western-trained doctors. Indeed, some Tibetan physicians are intentionally undertaking Western medical training.
From the perspective of our hope of understanding the traditional Tibetan medical system, this trend may seem to be a problem, raising the question of 'contamination' of the traditional Tibetan system by Western influences -- but the Tibetan physicians certainly have the right, and indeed the responsibility, to learn and use whatever methods they believe will best serve the health of their patients.
This influence from Western medicine affects not only the classification of disease entities but also diagnostic procedures. For example, asking patients to describe their symptoms was not traditionally part of the Tibetan physician's diagnostic examination. Pulse analysis and inspection of the urine, along with the visual appearance of the patient and other signs, gave the physician all the information needed to select an appropriate treatment. However, lately Tibetan physicians who have been working with Western patients are asking them to describe their problems.
However, this cross-training of physicians in both systems also presents an opportunity. One very powerful way of studying the relationship between the Western and Tibetan classification systems for medical problems is to interview physicians who have trained in both systems.
We may also see the growing influence on Tibetan medical practice from Western medicine, and from Tibetan physicians simply living in a Western or Westernized society, as a challenge. If we hope to study the Tibetan system of medicine while it retains most of its traditional form, we must quickly marshal the resources and energy required for doing so. As we proceed, we must also be alert for distorting effects of the research itself, such as a tendency to consider the specific treatments that have been validated experimentally as superior to other treatments, even though the other treatments have not been tested.
top of page
Tibetan Medicine at Dharma Haven
Natural Healing Wisdom for the Modern World
Medical Research, Tibetan Style
Resources on the Web
to Validity of Research Design - Barbara Ohlund & Alex Yu
Jürgen C. Aschoff's Tibetan Medicine - Tibetische Medizin is an annotated bibliography of books and articles published in all languages except Tibetan from the year 1789 up to 1995, including an extensive collection of reports of Western medical research
top of page
Your Comments and Suggestions
Published on August 12, 2002
Revised on September 3, 2002
Copyright © 2002 Dharma Haven
top of page