Truth or probability? A proper scientific identity for homeopathy

Table of contents

1. Introduction

The Method and Validation Committee of the Dutch Society of Homeopathic Physicians (VHAN) has been evaluating homeopathic remedies for several years by means of consensus meetings of colleagues. In this process we are confronted with essential epistemological questions. This paper is intended to promote discussion about a scientific identity that suits the homeopathic profession. Our premise is that homeopathy is far from perfect and that there is no clear or unanimous program for the perfection of the method. Our premise is that the rationale and method of homeopathy lack consistency and validity in a number of important respects, and that there is no clear or unanimous program for remedying these weaknesses.' I think it is worth being that much more explicit. The development of a proper scientific identity must include the search for useful starting-points, or key concepts, from within a coherent philosophical framework. It is possible that these do not lie within the conceptual framework of conventional medical but elsewhere. This is an unknown terrain for us as medical doctors. So we invite anybody whom these matters concern, especially from other professions, to join this discussion.

According to Prof. J.P. Vandenbroucke, the scientific evidence for homeopathy is as strong as for any conventional therapy. Because of the implausibility if the mechanism of action, doubts arise about the validity of even the most positive results of Randomised Clinical Trials (RCT), and the appropriateness of the RCT for investigating homeopathy.' -The process leading to the acceptance has come to a deadlock. As a result, the process of homeopathy has become deadlocked because thus far the implausible mechanism of action was no problem in this respect. Kleijnen states that it is not useful to do more randomised trials(1.).

Homeopathic physicians gather knowledge and they use it without much reflection about this knowledge. It works and that's it. In the last 15 years we have seen new developments; theoretical concepts and all kinds of provings are used to find new remedies. We see hypotheses that remain implicit and thus are not validated. There is a tendency to incorporate these hypotheses blindly in our methodology.

The research we practise is mostly modelled on conventional research, like the RCT. It seems, however, that the RCT does not add much to the efficacy of our method. The RCT does not do much, however, to assist the development and effectiveness of our therapeutic method. The methodology of the randomised trial imposes several limitations, such as the exclusion of co-morbidity. In the homeopathic methodology co-morbidity is the key to our remedy selection; the more we know, the more precise the choice of the remedy.

It is not very satisfactory that the methodology and the applied knowledge used by the homeopathic physician depends so much upon observation and intuition. It must be possible to find or develop instruments that make our method more evidence-based instead of opinion-based. Maybe these instruments can be found in other fields of science. In ICT e.g. we use fuzzy logic and knowledge-systems to derive successful strategies from experience. Homeopathy has much to do with experience and disciplined observation.

The acceptance of homeopathy as a science is particularly difficult because we use a different theoretical approach. It seems that we operate under a different paradigm. We handle this slightly awkwardly: we say that every choice of a remedy is unique, but that is something hard to digest for a scientist. In fact this uniqueness is a misconception, we develop consensus about the type of patient that needs a certain remedy. This 'type of patient' constitutes a class of qualities that contains mental and physiological entities apart from pathophysiological diagnoses.

We think homeopathy can be made much more efficacious by a scientific approach. Probably an adaptation and an extension of the conventional instruments in medical science is necessary. This leads to a proper scientific identity, like general practice in medicine has developed a proper scientific identity. E.g. consensus and operational standards have become incorporated alongside randomised trials(2.).

This paper tries to give an overview and keep all things together. Everything has to be worked out more exactly, we just hope to indicate a way to go. The main problem we expect is how to introduce science and still safeguard a holistic approach. We expect that keeping all variables in scope will hinder us in detecting the truth. Things will only become more probable.

1.1. Instructions for the reader

We opted for a circular structure in this paper, where the problems are dealt with following concentric circles. From general remarks on we work towards a homeopathic elaboration of what we think might constitute the homeopathic scientific identity. So you can see the same subject returning in later chapters enlightening more of the homeopathic aspects. In the chapter 'Consequence' we seek for solutions illustrated by some developments already in process. In each following chapter a homeopathic paradigm becomes more tangible. In the chapter 'Finally' we make some practical remarks and we conclude with our recommendations.

2. Summary

We have analysed the place of homeopathy in the scientific field. Homeopathy as a scientific method is far from perfect. Two centuries ago medicine was divided into a more reductionistic approach (now conventional medicine) and homeopathy, which persisted in the study of the whole being. When homeopathic physicians and their conventional colleagues communicate a difference in paradigms becomes apparent. There are essential differences in the expected outcome of homeopathy and conventional medicine. Homeopathic knowledge consists merely of practice experience, theory is scarce. Homeopathic theory is mainly system-theory. A crucial question in our method is the causal relationship between remedy and cure. The RCT is not very suitable to improve homeopathy. Accurate case-description, knowledge of the spontaneous course of illness and comparison with this and consensus give an indication of the causal relationship between remedy and effect. Within family practice this procedure becomes more appraised. In other fields of science, like system-theory, psychotherapy and neural networks, casuistry is the main element of knowledge. Knowledge from experience becomes accepted beside statistical knowledge. The existing scientific arsenal must be suitable for further development of our method, although some adjustments may be needed.

We do not start from scratch. During the last two centuries a methodology has emerged that requires assessment. Methodology and instruments we use have their roots in the nineteenth century. Hypotheses are not enough validated. Materia medica and the repertory based on it need validating and structural reform. The value of remedy-proving and different potencies is as yet unclear. Methods of investigation like single case design, fundamental research and cost-effectiveness research must get more attention. New scientific knowledge and new tools (among which information theory seems to present great opportunities) must be incorporated.

Much of this is already going on. Information theory has shown shortcomings in our instruments and indicated the crucial role of semantics. System analysis reveals our algorithms lacking explicitness. We recount some themes that are discussed among professionals. Actions of homeopathic physicians are more and more recorded on digital media and therefore accessible for evaluation. We are validating materia medica with structured consensus meetings.

This paper is meant to instigate a scientific policy plan. The profession must get on its way to a proper scientific identity and make this clear to the outside world. 'Recognition' is a comparative notion, but has much influence on funding of research. Funding out of the sales of homeopathic medicines seems still far away. We will have to make use of the fact that homeopathy meets world-wide consensus so that tasks can be spread all over the world.Our conclusion is that homeopathy is not up to date in a scientific respect. The instruments are outdated. Reflection of truth in our observations and hypotheses is possible and necessary. Sifting of untrue insights must generate a more effective method. A more effective method will convince more people. The responsibility for this lays in the hands of all homeopathic practitioners, who have to develop more team-spirit to gather the relevant data.

3. Starting-point for the scientific identity

History shows that homeopathy always had its own scientific identity. This identity was set in the nineteenth century. Later developments in science have been applied to homeopathy but not as an integrated procedure to develop the method. The randomised trial has been exclusively used to prove the efficacy. In the twentieth century some new concepts are introduced like paradigm and placebo effect. Precise recording of premises and hypotheses in research became a standard procedure. To make this feasible a tendency to reductionism arose. In the same line of thought a preference for statistics and a repugnance to casuistry was logical. Here homeopathy failed to come along. >/p>

Since the introduction of the computer new trends developed: on one hand the development of software needed meticulous analysis and description of procedures and systems, on the other hand it became possible to discover patterns in heaps of data. About the same time first line medicine became fully-fledged. Here the limitations of statistics became visible. Casuistry and common knowledge were revalued. These latest developments offered new points of contact for homeopathy.

In this chapter we arrange the starting-points that are relevant for homeopathy's scientific identity. In the next chapter we make an inventory of the problematic areas. With some practical workouts we indicate a scope for the future. It may seem a bit fragmentary and wide-ranging, but we want to keep an open mind to as many options as possible.

3.1. Goal of scientific research

Anybody will agree that scientific research is meant to distinguish sense from nonsense. There is however a difference in staging. The homeopathic physician supposes that homeopathy is effective and wants to improve the method. The critical outsider wants to see proof of the efficacy. This is an chicken and egg situation: an imperfect method cannot give optimal result in randomised trials and real interest from outsiders is lacking because of the imperfect evidence.

When asked who benefits from scientific research we regard first of all the directly involved parties. That the patient is involved goes without saying. The call for evidence however comes mostly from the conventional medical profession. Patients do not seem as keen for proof because they expect something from homeopathy that so-far conventional medicine could not deliver. Especially homeopathic doctors and the homeopathic pharmaceutical industries have direct interest in scientific research. They have to optimise the available resources and time. Remedies and methods that that are not effective counteract the progress of the method.

This paper is written from a homeopathic view. From this standpoint the goal of scientific research in homeopathy is, in the first place, more knowledge about remedies and their application. With this objective we start with an analysis of the historically originated characteristics of the method.

3.2. Homeopathy and development of science

When homeopathy was discovered, just before 1800, more radical changes took place in a badly performing medical profession, like the development of physiology, researching the function of organs by vivisection of animals. Somewhat later chemistry arose. By researching the function of separate organs a course quite different from homeopathy was taken. Homeopathy regards the intact system as a whole and tries to draw reproducible conclusions. This diversion of courses has caused a fanatic religious quarrel: holism against reductionism. Already at the end of the nineteenth century it appears that reductionism is not as successful as expected. Nor physiology, nor chemistry, nor randomised trials lead to consistent and fully reproducible knowledge.(3) Homeopathy was based on experiments in a time that medicine was strongly opinion-based. Experiments with medicines on people have later become in vogue as phase 2 research in conventional medicine. Apart from this homeopathy took its own course: meticulous clinical observation, structured and lengthy consultation, detailed reports and exchange of casuistry have been the professional standard from the beginning.

A specific aspect of the homeopathic method is knowledge retrieval. The large amount of data has to be handled during consultation. A distinguishing phenomenon in homeopathic practice is the ample use of books, currently also in digitalised format. Arranging data and making them accessible is a core-activity in homeopathy. The same goes for other fields of science, like jurisprudence.

3.3. Paradigm

If homeopathy is based on a different paradigm is still debatable. Conventional medicine denies it, but isn't that the essence of the notion paradigm? Followers of a competing paradigm do not like the possibility of another paradigm. Indications for the existence of a deflecting paradigm come forward in the sparse debates between conventional medicine and homeopathy. An example was the discussion around the academic thesis of dr. Martien Brands.(4) This thesis is about a semiotic analysis of consultations in conventional medicine, homeopathy and Chinese medicine. Stomach pain is put as an example. Comparing homeopathy and conventional medicine the author concludes that symptoms like 'burning pain', 'perspiration of feet' and 'fear of burglars' - important symptoms in homeopathy - do not influence diagnosis nor treatment in conventional medicine. All opponents without homeopathic experience disagreed with this assumption. They stated that such symptoms are also used in conventional medicine. Nevertheless anyone with conventional as well as homeopathic experience will share the view of Martien Brands. Such symptoms can be of even greater importance when it comes to choosing the right remedy. Only the experience with homeopathic remedies can lead to such a conviction.(5) When a physician makes the acquaintance with homeopathy he experiences a culture shock. Beforehand he would not consider 'fear of burglars' as more important than the diagnosis gastric ulcer regarding the choice of a therapy.

A different paradigm can interfere with the communication between homeopathic practitioners and the scientists who advise them. The scientist who did not experience the culture shock cannot understand why a certain instrument doesn't work or why other instruments are needed. The question of paradigms is so essential in respect with scientific identity that it appears necessary to elaborate this subject, e.g. into an academic thesis.

3.4. Expected outcome

The expected outcome of a homeopathic medicine is hard to define beforehand. As a whole it comes to better functionality and well being. Imagine that in the example mentioned above (stomach pain) the well being improves (possibly expressed in less fear of burglars), but the stomach pain does not improve within the time set for the trial. According to homeopathic standards this is a better result than improvement of the stomach pain with equal fear of burglars. One of the rules in homeopathy states that mental symptoms should improve before physical symptoms. Therefore 'bad homeopathy' can be necessary for good results in randomised trials. In other words: 'stomach pain' is a bad substitute for the real outcome, better functionality and well-being.(6) To avoid this we should exclude patients with 'deeper' pathology, but in real life those of all patients benefit most from homeopathy.

Another problem is the fact that only an estimated 50 homeopathic remedies are sufficiently tested to make a state-of-the-art application possible. The other remedies are more or far more difficult to find because of the lack of knowledge. Based on the premise that a homeopathic remedy must suit the type of person we may expect to need many different remedies to cure everybody. If we take the number of needed remedies to be 500, we must conclude that in only 10% of the cases a state-of-the-art prescription is possible. Only in these cases we may expect the best results, but even then, we do not know if the complaints we measure will subside within the time set for the trial. If the remedy does not fit the patient totally, only a part of the symptoms will subside. Chance will decide if the symptom under investigation is among them.

Regarding these limitations it's amazing that the results of randomised trials are not so bad after all. But when we have to deliver more and better proof to overcome the theoretical objections against homeopathy we must be very careful. It is likely that better results are impossible as long as the method has the weaknesses mentioned above and later on in this paper. In fact many of the prescriptions we make are placebo-prescriptions because we chose the wrong remedy due to insufficient knowledge about remedies and methods. Better homeopathy will give better results in randomised trials.

To improve the method we need instruments that reduce reality in the least possible way. We need to preserve all possible aspects of each case. Can that be done and how?

3.5. Knowledge from each case

The problem with reductionism is also known in conventional medicine. The outcome of randomised trials appears less valid in daily practice. The efficacy of a physician in diagnostic respect increases with each year of experience. Experiential knowledge plays an important role in daily medical practice, maybe more important than knowledge from epidemiological research. The same goes for homeopathy. The problem is how to make 'hard' facts out of this knowledge.

Experiential knowledge shows that each individual case can deliver knowledge. Homeopathic knowledge arises from large amounts of casuistry. In this casuistry successful patterns become visible regarding techniques of case taking, evaluation of data and materia medica. The way this knowledge comes forth and how to improve this is hardly a subject for investigation. The description of cases is thus far mainly qualitative and less quantitative. In the nineteenth century a standardisation in case-description developed that is still in use. Frequently homeopathic casuistry contains implicit hypotheses that are not validated or hypotheses that meet no consensus.

3.6. Post aut propter

Before accepting a case as a source of knowledge a selection-process takes place. There has to be an apparent causal relationship between the remedy and an effect. The judgement about this causal relationship depends on 'knowledge about the world', an expectation about the chance of spontaneous recovery and the spontaneous course of illness based on experience in daily practice. Such knowledge is often not founded on statistical data. Furthermore there is the role of placebo-effects. Nevertheless our judgement about the causal relationship between a cure and a therapeutic process contains a quantitative element. Most practitioners will see the recovery of an acute illness as spontaneous. But when a patient suffering from emphysema experiences a substantial and lasting improvement in functionality after a certain therapy we will suspect strongly a causal relationship. In daily practice each practitioner makes an (implicit) estimation of an effect, based on aspects that are described in chapter ('Causality'). This estimation has to be made more explicit.

3.7. Hypotheses

Homeopathy always stays close to the experiment, but we cannot avoid making hypotheses to arrange our observations. The hypotheses in homeopathy rather serve to facilitate the process than to explain things. Notions like 'self-recovery', Hering's rule (describing patterns in recovery),(7) theory of miasms etc. serve to arrange our casuistry and to grasp the process that we presumably influence. Hypotheses can lead to confirmation bias (see below). Regularly a hypothesis is used to substantiate a case rather than the reverse. In that case a false hypothesis is a weak foundation for a construction of new theories.

3.8. Truth and consensus

There is a gradual scale between the opinion of one person and a 'truth' accepted by everyone. Consensus therefore is no definite notion. Our profession relies greatly on consensus, but we are not aware of the mechanisms that constitute consensus and of the relation between consensus and truth. Bloodletting met general consensus for a long time. Retrospectively we must conclude that sometimes the interest of the profession is better served by consensus than the interest of the patient. Nevertheless consensus is rediscovered by conventional medicine. An example is the composition of professional standards for GP's in Holland. This is done under regulations to preserve independence and integrity of consensus. This consensus however is not always put into practice. Whether this is due to improper consensus or to unsound premises backing the consensus is a salient question. It is possible that a consensus does not reckon with the circumstances of the patient. Circumstances known to the GP, but not taken into account in randomised trials that constitute the consensus.

3.9. System theory

Apart from truth there are also social aspects of learning organisations connected with consensus. The group-process must produce additional value to the individual expertise and experience of the members of the group. Senge illustrates this in his 'fifth discipline'.(8) The system-theoretical approach seems to fit homeopathy because of the phenomenological character of the method. Homeopathic theory is mainly system-theory, e.g. Hering's rule that describes the desired course of the cure. The onset of homeopathic knowledge also forms a system. We search for successful strategies based on practice-experience. The group-process is essential in this respect because there is much esteem for common opinion in the profession.

3.10. Confirmation bias

Confirmation bias stands for selectivity in thinking or observing, where things that agree with one's belief are noticed and things that do not agree with this belief remain unnoticed . More intellectual exercise is needed to deny an established fact than to recognise it as such.(10)

Concerning homeopathy consensus has been established about its principles and materia medica surviving generations and different cultures, without any enforcement. Apart from that there are groups within the profession that have ideas not shared by other groups, or ideas that are deserted in a next generation. Interests and group-processes may play a role in this phenomenon, like material interests, insecurity and emotional dependency towards gurus. These mechanisms must be born in mind to avoid incorrect consensus. There must be a constant readiness to falsify all consensus.

It seems that conventional medicine and homeopathy can learn from each other. This is mutual. Awareness of prejudices is not a strongly developed skill in homeopathy. Psychiatry e.g. has a much stronger tradition in this respect.(11) Gathering casuistry we must always suspect confirmation bias. One of the means to avoid this is the quantification of data.

3.11. Truth can be measured

Apart from moral aspects of independence, consensus has quantifiable aspects. Such as numbers of participants in the process, measures of objectivity and the proportion of parts that meet a great deal of consensus and parts that meet little consensus.

All the data that constitute the consensus need to be assessed. There are many things that can be measured, e.g. in the materia medica the prevalence of symptoms. The interpretation is something else. Before we draw conclusions from casuistry we need protocols for data-collection. We need to define cure. Symptoms show great diversity,(12) semantic problems increase this diversity. Statistical entities like sensitivity and specificity relate to the population as a whole, of which we know very little.

Theoretically there are many clues to measurements. Basically the homeopathic method consists of semantics and statistics. This needs clarification from a rather abstract point of view:
Homeopathy deals with symptoms of a patient. Any recognisable symptom is given a place in the semantic space that constitutes the knowledge-domain of the homeopathic physician. The symptom gets a statistic measure, the rarity of the symptom. This is expressed in paragraph 153 of Hahnemann's Organon,(13) the rarer the symptom, the greater the information-value. The symptom can be an element of different collections in the knowledge-domain constituting the materia medica. By adding symptoms we decrease the possible number of remedies. The information-value of each symptom is as yet estimated, but in principle it can be measured. A complicating factor is semantics. We have need of a set of co-ordinates that can express the semantic space in numbers. Such sets are in progress as indicated by semantic networks. Homeopathy as a mathematical process seems less utopic.

3.12. Logic

Many people reading Hahnemann are struck by his logical reasoning considering his era. Such logical reasoning deserves more attention in the scientific approach of homeopathy. With reasoning it is possible to discard some phenomena in homeopathy as unlikely. Unlikely does not mean untrue, but it deserves a more critical attitude. Examples are: miasmas, distinction between acute and chronic remedies, dream-provings and the dogma of signatures. One can also question the validity of provings: how is it possible to evoke symptoms with non-substantial doses in a small group of subjects, when we assume that the action of a remedy is restricted to a rather specific person? The placebo effect was unknown when provings were established.

3.13. Hard data

We are exploring ways to turn 'soft' data into 'hard' data, in other words convert them from qualitative to quantitative. The information offered by the vast amount of data from casuistry is relatively soft due to problems like semantics, intermingling with placebo effect and interpersonal variability. On the other hand these data reflect day-to-day practice and not some experimental situation.
In other fields of science working with day-to-day practice is also common practice, like in neural networks: these are machines with hardware or software simulating neurones. They are capable of detecting successful procedures out of a vast amount of cases. This appears to be a very objective procedure with measurable results. The use of these machines is still limited to concrete subjects like process-technology. Semantic problems prohibit the use in homeopathy and these problems cannot be solved at short notice.
It must be possible to express these techniques in formulas and algorithms and to find parallels with human selection of successful strategies. Maybe the techniques of a neural network can be combined with the associative powers of the human mind.

3.14. Art and knowledge

We start with a parable:(14)
Once upon a time there was a centipede dancing magnificently with all her legs. When she danced all the animals in the forest came to watch. Every animal was impressed by this magnificent dance, all but one. The toad did not like to see the centipede dance and he thought: "How can I stop the centipede from dancing?". He could not just say that he did not enjoy the dance. He could neither say that he danced better, that would sound ridiculous. Then he hatched a cunning plan. He wrote a letter to the centipede: "oh, unexcelled centipede!" he wrote, "I am a most dedicated admirer of your dancing art. And I should like to know how you do it. Is it so that you lift at first your left leg number 228 and then your right leg number 59? Or do you start dancing by lifting right leg number 26 before lifting right leg number 449? I am most interested in your answer. Greetings, the toad".
The centipede never danced afterwards and is still thinking about the answer.

The moral of this story is that science can kill art. This doesn't imply that science and art are inimical, otherwise art-schools would not exist. Art and science are complements. Medicine cannot exist without art, maybe this goes even more for homeopathy. Does a holistic view imply a more artistic approach?

3.15. Pattern recognition

Art in medicine is many times some sort of pattern recognition. This calls for further investigation. Different diagnostic strategies can be used to find the right homeopathic remedy. These strategies can be compared with the process in the consulting room of the GP.(15)

Frequently used strategies are:

• Pattern recognition;
• The use of algorithms (decision tree)
• By sequential hypotheses (systematically discarding hypotheses)
• Random dragnet method

Pattern recognition, as a strategy is an important tool for the experienced physician. It enables instant recognition of a diagnosis or illness otherwise hard to recognise.

3.16. Standardisation?

It's fatal to theorise without facts (Sherlock Holmes). Facts are abundant in homeopathy, theories scarce. It would give us considerable new possibilities if the mechanism of action were plausible. Knowing the mechanism would e.g. help us to choose the remedy by deduction.
Still we must make some annotations here. The strength of homeopathy lies in the holistic approach. Just that fills the gap that is caused by reductionism. Holism is necessarily an inductive approach where theory is rather difficult to establish.(16) It is hard to imagine that the theory we are looking for is a standard pharmacological theory because of the infinitesimal dilutions. In our profession there is hardly any knowledge concerning natural sciences and there is no structural relationship with these sciences. We are not even sure if the theory that we are looking for can be found within the conventional scientific paradigm.

The existing scientific sources that are momentarily in range are chaos-theory,(17) electromagnetic changes,(18) resonance(19) and water-structures.(20) The gap between these theories and the functioning of living creatures is still very large. We want to stress here that any theory must be assessed by the facts (our casuistry). The best theory must win by falsification. There must be an interaction between those who make the theory and those who test it. Theories that cannot be falsified must be banned.

3.18. A different way?

Searching for a proper scientific identity we discover the need for better definition and evaluation of premises, assessment of already existing knowledge about homeopathic medicines, enlargement of the knowledge about current remedies, possibly discovery of new remedies and more insight in the process of pattern recognition. We stress that effectiveness (in daily practice) must remain our goal.
Partly we can join conventional medicine. Several fields of medicine (like mental health care) need their own instruments and there are developments in that direction.(2) We can join this development and we have something to offer too. We have a strong tradition in consensus.
We have made some critical remarks concerning randomised trials. Still we acknowledge that prospective, placebo controlled and randomised trials have the highest internal validity. If it fits our goals this type of research is to be preferred. The second important aspect of the homeopathic scientific identity is knowledge retrieval. Other scientific fields like jurisdiction, linguistics and ICT have similar problems. So it seems wise to closely monitor developments in these fields. One aspect needs special attention: nearly all of our knowledge is based on casuistry. We assume a causal relation between medicine and cure. If this assumption is not correct the whole building collapses. We conclude that we do not have to re-invent science. We must look for the right tools in the existing toolbox. Basically homeopathy can still join existing science. There is a possibility however, if there is indeed a conflict in paradigm, that we must search in unusual places or adjust the usual tools.

4. Inventory of problems

Following a general description of aspects concerning a proper scientific identity we come to problems known in the homeopathic field. We search for quantifiable elements and try to establish a certain amount of probability with those elements. Clinical experience still plays the leading role in this.
Apart from questions about quantity and validity we wonder what sort of problems a legacy of two centuries can pose: e.g. mistakes by copying essential material by hand and gathering of data by out-of-date techniques.
Each subject in this chapter deserves much more study. In the next chapter we give some hints about discussions already taking place. In order to keep an overview we do not elaborate things here. We see this overview as a guide to keep the bottom line in sight in future investigations.

4.1. Definition of cure

In chapter 'Expected outcome' We stated that the goal of homeopathy is wellbeing and better functionality. A practical definition of cure is still hard to define. The Society of Homeopathic Physicians in Holland (VHAN) has reached consensus about an operational definition in the past years. This definition is based on some premises in homeopathy that are not yet fully validated, but it helps to define successful cases that can be used as a start in the process of induction. This consensus however is not definitive.

This is the Dutch consensus:(21)
On October 4 1996 a group of 80 physicians, members of the VHAN, discussed criteria we can use to indicate a case as being successful. Most positive results are indicated by 'better general wellbeing' (including emotional state) and 'energy'. Followed by amelioration of the main complaint provided there is a development following Hering's rule. Then bodily functions like sleep appetite, stool, temperature, menses and sexuality. And lastly clinical parameters like physical examination, laboratory findings and the use of allopathic drugs. The occurrence of an initial aggravation is an indication for an effect, but not a criterion for cure.
There is a difference between acute and chronic cases. Spontaneous recovery is most likely in acute cases. In acute cases the amelioration of the main complaint is more important regarding the assessment of the result. The state of health is expressed in a 9-point scale (-4 to+4),(22) closely related to the 'Glasgow Homeopathic Scale', but slightly more precisely defined.

4.2. The quality of the case

Regarding the importance of casuistry as to the development of our profession we must strive for the utmost quality in this respect. Every doctor who shares cases with colleagues has to be aware of his scientific responsibility. The most important quality-indications of a case are descriptions of causality and premises. In present practice we take existing methods and knowledge for granted. In the next chapter we deal with old casuistry. Silently we adapted new techniques, like 'measuring'. Regarding casuistry in the future we must have a clear indication of quality.

4.2.1. Causality

The causal relation between medicine and cure is hard to establish but essential. Only those effects that can be attributed to a medicine can play a role in the evaluation and improvement of the method. In Holland homeopathic physicians mostly treat chronic cases. Spontaneous recovery is infrequent in those cases, but variations may occur (e.g. in rheumatic illness and M.S.). A physician can get an idea about the action of a medicine in a case by reading a full and adequate description. It would be desirable to standardise description of cases. Each case-description should include a paragraph containing:

• Duration of complaints before starting treatment
• Severity of complaints, also related to the context of the patient
• Description of the improvement, as quantitative as possible
• Duration of follow-up, frequency of medication
• Reasons to repeat the remedy
• Change in clinical parameters and use of medication
• Case-mix: external factors influencing results

Assessment of results could be much more accurate if we could use epidemiological data about spontaneous recovery of illnesses. These data can also be mentioned in the description of the case. It is at this moment hard to say anything meaningful about the role of placebos. The notion that one-third of all patients responds to placebo is a definite misconception.(23)

4.2.2. Case-handling

When we see descriptions and evaluations of cases a great variety of case handling appears. Existing rules like § 153 and Hering's rule meet different interpretations. Some practitioners seldom change remedies, others frequently. Is there a rational explanation? How do we research the effectiveness of different methods?
Apart from this the question is what we want to treat. This question can be different in each case, but it must be put in the onset of each treatment. Do we wish to treat the complaint at hand or do we wish to improve wellbeing and is this always possible? What can we expect considering the illness at hand and the vitality?

4.2.3. 'Measuring'-techniques

Some homeopathic physicians frequently use 'measuring'-techniques. Very little is published concerning this issue. Measuring suggests exactness and reproducibility, but that is as yet far from proven. There are indications that measuring-techniques resemble commuting. In our society commuting and intuition has a bad reputation. This may be unjust, but on the other hand we may not present these techniques as exact instruments. Practitioners who use these techniques must in some way account for it and try to assess these methods. Case descriptions must be clear about application of 'measuring'-techniques.

4.2.4. Validating hypotheses, falsification

The need to be more conscious of hypotheses has been mentioned before. The homeopathic physician must be critical towards hypotheses. A case must serve to back a hypothesis and not the other way round. Some very doubtful hypotheses are slipping into homeopathy like the dogma of signatures. The profession must develop a critical attitude towards hypotheses and gather cases that contradict them. It would be wise to make a public list of all hypotheses that are just in homeopathy with a database of cases that contradict the same. This will not discard every hypothesis but put it in the right place. If we for instance collect all the cases that contradict Hering's rule we know better what it's worth.

4.2.5. Peer review

Assessing the quality of a case has to be standardised. Its needs a permanent committee. Such a committee should assess every case that is offered for publication, in journals in courses or elsewhere. The committee can provide cases with a hallmark, showing quality aspects as mentioned before under 'Causality'.

4.3. Instruments

Our current instruments, especially materia medica and repertories, are out of date and poorly standardised. Semantic problems make our materia medica hardly accessible for electronic searching during the consultation. The search is nor precise (only that what you are looking for), nor complete (all you are looking for). In the repertories the semantic problem is partly solved by putting synonyms under one heading. The greatest problem of the repertories however is the transcription from the materia medica by hand implicating many mistakes. Using modern techniques we probably would have developed other instruments instead of the repertory. Nevertheless the repertory still functions adequately, due to the profound knowledge every homeopathic physician has regarding this book. Because of this knowledge the information in the repertory recalls many associations. Repertory problems will be dealt with more extensively in next chapter.
The basis of the repertory is a materia medica from a nebulous origin that is also poorly standardised. This also has to be reconsidered.

4.4. Validating materia medica

At the end of a case (successful or not) we assess the materia medica on which the choice of the remedy was based. This is a circular process. In Holland there is enthusiastic exchange of experience among colleagues. There are regular meetings to validate materia medica. On each meeting two remedies are assessed. We compare cases to find similarities. In the process we discover difficulties of various nature, like causality, semantics and confirmation bias. Evaluating these data we consider the similarity (and dissimilarity) between symptoms and diagnostic instruments, mainly with regard to statistics. When we overcome these difficulties we have more insight into the advantages and difficulties arising from large computerised databases.

4.4.1. Semantics

Semantics cause serious problems validating materia medica. Many professionals use different words for the same concept, one professional can even use different words for the same concept in his own cases. How can we compare casuistry then? Even methodology must be semantically analysed: What do we mean by 'classical homeopathy' and what do we understand by a proving? Another disadvantage of such terms is the obscurity to the outside world. Communication with others would benefit from standardised terminology.(24)

4.4.2. Registration

We are gathering successful cases that meet the Dutch consensus. Registration of cases becomes more popular. Now we need tools to make the best possible use of these data. The existing computer-programs must comply with the need for optimal registration, neither too much nor too few data.

Registration can serve many purposes,(25) but one can ask if all purposes have to be served in the same time. The more data we gather, the more time-consuming it gets. Then there is the risk that less colleagues co-operate. The most appropriate goals are:

• Validating materia medica
• Making inventories of results concerning diagnoses
• Effective practice (Book-keeping, billing, correspondence)

There are data that need constant registration and data that do not need this. Its obvious that the results concerning prescribed medicines needs constant registration in order to make validation of materia medica possible. Results concerning diagnoses can by gathered by some practices during a limited period, this is sufficient to give a reliable outcome. The same goes for initial aggravations, once we know its prevalence, it's not likely to change in the near future.
After registration in each practice the data must be gathered. This presents specific problems like standardisation and protection of privacy. The following step is interpretation of data. The most interesting is the quantitative aspect: The prevalence of each symptom leading to sensitivity and specificity. More about this in the next chapter.

4.5. Databases

To gather as much data as possible in a sensible way we need a carefully built international database of clinical outcome and provings. Such a database is the Delphi project. Other initiatives in this field come from HomInt and ECHAMP.

We need an international protocol for the data regulating things like:

• Criteria for underlying provings and clinical material
• Criteria for description of cases as mentioned in the previous chapter
• Standardised description of sources, numbers of participants in a proving or patients, verifications, authors
• Linguistic directives
• Schemes for re-evaluation and priorities
• Criteria and scales for the quality of the information

4.6. System analysis

Hahnemann and Kent extensively describe the process leading to the choice of a homeopathic remedy. These descriptions are mainly qualitative. Algorithms in homeopathic software prove that the rules can be made quantitative. The most important rule in homeopathy, par. 153 in Hahnemann's Organon, can be translated in terms of sensitivity and specificity. The computer-programs enable us, by these algorithms, to repertorise with special regard to small rubrics in the repertory or to small remedies. Another example of the interference of ICT with homeopathy we see in expert systems: using if-then rules we can detect very specific combinations of symptoms leading to a specific remedy.
The system-analysis for computerised aid in homeopathy is so far a purely commercial business. That is not good for the development of homeopathy.

4.7. Remedy provings

Hahnemann's description of provings was accurate regarding his era. But things have changed. Conventional Phase-I research has led to rules to protect the subjects of experiments. Hahnemann was not aware of the placebo effect as far as we know. To solve this problem with placebo control seems contradictory to the essence of the proving, i.e. describing every symptom of the subject and not only symptoms that occur significantly more compared to the control-group. In that case only intoxication-symptoms remain. Solutions may be found in the single-case-study design.(26) Even then it is difficult to discern which symptoms are relevant.(27)
The remedy proving has to be validated, e.g. by repeating old provings and conducting simultaneously the same proving on different locations.
Critical review of the process regarding provings by outsiders must be welcomed. New knowledge about suggestion, projection and other psychological processes in subjects and conductors of provings must be incorporated.

4.8. Different potencies

Hahnemann's stunning discovery that potencies may enhance the curative strength of a medicine is the greatest stumbling block for conventional medicine. It has also caused endless debates among homeopaths. This debate has not led to a conclusion as yet. The production of unnecessary potencies is a waste of time and an obstacle in comparing cases. We need to investigate this and reach for consensus.

4.9. Single cases study

Sometimes we learn from coincidence: the practitioner or the pharmacist makes a mistake in prescribing respectively in delivery, someone takes the remedy prescribed for another member of the family. The outcome of such cases can be of great value because of unexpected result. It's interesting to make a special heading for such cases in the databases mentioned earlier. Retrospective analysis of these cases can give interesting outcomes.
The classical prospective single-case study must still prove its usefulness in homeopathy. It is possible that the carry-over effect of homeopathy can pose problems in interpretations. Maybe a special design of such a study is necessary. If the single-case study is feasible we can use it for a large scale of questions such as: difference between potencies, different effects of different remedies in the same patient etc..

4.10. Observational study

Observational studies are more useful for the development of homeopathy than randomised trials, particularly to assess the applicability of homeopathy in daily practice,(28, 29) especially in relation to conventional medicine. It may be a useful method to improve the method as such by comparing different hypotheses.

4.11. Fundamental research

For fundamental research we need co-operation with other disciplines. This is not easy to obtain, at least not in Holland. Even the research of van Wijk did not soften the rigid rejection of the Dutch universities.(30) When it comes to prove homeopathy we do not expect much from research that has been conducted by or paid for by the homeopathic community. We plead for fundamental research that serves the goal stated in this paper: improve homeopathy.

4.12. Cost-effectiveness

When proving efficacy is a dead end we would do better to focus on social relevance like cost-effectiveness. As in observational studies it is useful to investigate the position of homeopathy in relation to conventional medicine. It can also be useful to compare different strategies in homeopathy. Cost-effectiveness studies are still not an essential part of medical scientific research and are not so well defined.(31)

4.13. Safety

Safety is taken for granted in homeopathy. History seems to indicate that the method is safe, all alarming reports thus far were related to herbal medicine. Randomised trial and other investigations show hardly any undesirable effect. Still we should have an independent registration of side effects.(32) We know that initial aggravations can be substantial. Are we sure that they are not dangerous and leave no permanent damage?

5. Consequence

So far we have made a survey of the existing problems. The consequence appears to become a circular process. While performing consensus meetings validating materia medica our committee encounters over and over deeper layers of the same problem: what is the source and the nature of our knowledge? We just went on gathering casuistry. This casuistry is based on old premises, the very same we want to investigate. Step by step we become more aware of these premises. This awareness leads to adjustment of procedures. At the moment we still don't feel competent to give very concrete advice as to the desired course of action. We have our experience but others may have experiences of a different nature. We limit ourselves by some global lines of thought and a survey of the preoccupation of the committee. This chapter addresses the homeopathic physician in the first place. If the culture shock has come by one can better understand these themes.

5.1. Rational approach

Since there is too much to be done, it comes to priorities. Every problem has its own urgency, difficulties and methodological problems. Urgency can be stated by factors like:

• How essential is the problem as to the premises of the method; e.g. can it make the method questionable?
• How much energy and efficacy is lost when a certain procedure is ineffective or is gained by the opposite?
• How much financial benefit can a certain approach, or the abandonment of an approach, give?

Some questions (like semantics, ponderation factors, provings, materia medica, repertory and pattern recognition) are yet under attention. Beside those we can think of several others:

1. Is the construction of new remedies out of theoretical premises justifiable?
2. Can the dogma of signatures be used to find new remedies?
3. Is the dream proving valid?
4. How many remedies do we need?
5. Is there ground to distinguish constitutional from acute remedies?

Dealing with such questions needs a firm protocol, containing elements such as: participants in the discussion, accurate definition of problems, literature, available research, quality of casuistry, epistemology, methods and falsification. Some questions are hard to answer, like the number of remedies we need. Still we need consistent and practical management. Some subjects need constant attention, like semantics and confirmation bias.

5.2. Clinical experience

Apart from provings and toxicology homeopathic materia medica is based on clinical experience. We expect clinical experience to become more important due to the possibilities offered by ICT. Most casuistry nowadays however stems from the nineteenth and early twentieth century. This is the casuistry dominating the literature we still use. Thereby we make a distinction between 'old' and 'new' casuistry.

5.2.1. Old casuistry

Old casuistry is described in most popular materia medica, like Clarke, Boericke, Allen and Hering. In old casuistry there is much attention for physical symptoms. Most resources are closely interwoven with each other and with the repertory of Kent. They stem from consensus among physicians of that time. Some natural selection has taken place over generations and cultures. In practice we take this information as being fairly reliable, but this might be because of the lack of better knowledge. New diagnostic and patho-physiological knowledge has not been incorporated.

5.2.2. New casuistry

New is defined here as casuistry from the twentieth century onwards, but possibly also undiscovered casuistry from the nineteenth century. The main common aspect of this casuistry is that it does not figure in the tools every homeopathic physician uses. This kind of information comes from a large variety of sources, not all of them reliable. From 1920 to 1970 homeopathy was somewhat obscure. From 1970 there has been an explosive growth, partly caused by New Age. This is important because there are indications that the method thereby has been infected with a nihilistic conception of science.(33) There has been (and maybe still is) great tolerance towards speculative theories and casuistry with a follow-up of merely several months, based on vague strategies and assumptions. Some signs of dogmatism are visible.
In new casuistry mental symptoms are (in general) more important than physical symptoms. This must be evaluated, because mental symptoms are more subjective and semantically more indistinct than physical symptoms.
Another phenomenon is the world-wide travelling tutor. These teachers present casuistry in a very attractive way, backed by videos taken from consultations. The persuasiveness of such material is great, but we must realise that it expresses the experience of one person. One may get the impression that some of these teachers have expert knowledge of hundreds of remedies by experience. On statistical grounds this is unlikely.(34) Since 1990 there is a change. Professionals become more self-confident and exchanging own experiences becomes more important.

5.3. Materia medica

Homeopathic materia medica consists of provings, intoxication-symptoms and clinical experience. Fundamental questions to be asked about these sources therefore apply also to the materia medica. Some sources are reliable, some aren't.(35) Information from reliable sources is a major problem because most choices of remedies are based on a limited number of symptoms.
The portion of clinical experience in the materia medica causes a cumulative difference in the amount of knowledge for each remedy. A well-known remedy, with many symptoms know, is more frequently prescribed. Therefore knowledge about that remedy increases and so on. Casuistry is prose, not a collection of classified symptoms. When we compare a case with another case many similarities are not seen at first glance because synonyms are used or incorrectly seen because of homonyms. Classification of symptoms like the READ-classification is not yet available for all symptoms.

5.3.1. New remedies

We think different remedies are in a way related to each other. When we see symptoms indicating Natrium muriaticum and symptoms indicating Phosphorus many homeopathic physicians will prescribe Natrium phosphoricum. In a way the chemical salt is reflecting its constituents. But there are also specific symptoms not related to the constituents like the 'desire for fried eggs' in Natrium phosphoricum. When we assume that new remedies can be constructed by making all kinds of chemical salts of remedies we already know a large variety of remedies becomes available. Scholten has done this.(36) There are also other theories possible to construct new remedies.(37) We must stress that this construction of remedies is a theoretical procedure, where homeopathy is designed and proven as a phenomenological procedure. The procedure may be valid but has to be proved, which is not yet the case. There is not even a protocol to prove it. If we combine two remedies with 10 symptoms each we might construct 10,000 symptoms. Is it possible to prove that at all?
Apart from theoretical construction of new remedies there is a tendency to investigate all kinds of substances in regard to their usefulness as homeopathic remedies. This is a disturbing development when we consider that even the old remedies still need validation.

5.3.2. How many (new) remedies?

Most cases are treated with a top 50 of old remedies,(38) also in the randomised trials that indicate that homeopathy is more than a placebo-effect. To improve the results in randomised trials we can:

1. Validate the top 50 so that we make less mistakes with these remedies.
2. Further investigate the remaining 900 (or so) old remedies so that we can use them as effectively as the top 50.
3. Develop new remedies.

Is it too much to ask that new remedies must be as effective as the top 50 remedies? Is this possible when we develop more than 100 new remedies each year? We wonder if the people that develop the new remedies have a plan in this respect. We expect that 500 well-validated remedies suffice to make the method twice as effective.(39) To validate 500 remedies will still cause a enormous amount of work. If we use old or new remedies for this purpose this still has to be discussed.

5.4. Repertory

The materia medica is too vast to handle during consultation. Therefore an index to this materia medica is the standard tool. The most popular index is Kent's repertory, dated about 1905. The convenience of this repertory has altered homeopathic practice significantly. Instead of comparing full descriptions of remedies, we first put lists of remedies that belong to a limited amount of selected symptoms alongside each other in a spreadsheet. The result is a small amount of possible remedies that we can compare. The repertory encloses an enormous amount of data with many mistakes.

5.4.1. Technical problems

The repertory was made by hand. There have been several editions and revisions. In this process mistakes have been made, most of them typing errors. Kents' repertory contains some 650 remedies. They are reflected by abbreviations that resemble each other closely. One can expect that copying by hand causes mistakes. The 'Zeitschrift für Klassische Homöopathie' has dedicated a series of articles on this subject.(40)
Many rubrics in the repertory have sub-rubrics. The main rubrics do not contain all the remedies from the sub-rubrics. This is only just when the sub-rubric is a denial of the main rubric.

5.4.2. Methodological problems

The technical problems are easy to correct using the computer. Methodological problems are more difficult to handle. They are mostly semantic. Translating materia medica to repertory synonym and related words are united into one rubric in the repertory. This cannot be done consistently due to semantic problems. In spite of this there are still many synonym and closely related rubrics in the repertory. Related rubrics can contain very different sets of remedies.(41)
The loss of nuances in the repertory rubrics can be solved by hyperlinking the rubrics to the original texts in the materia medica . Before the computer, repertorisations were made by hand, but not in the exact way the computer does. Selections were made out of the different rubrics to save work. By doing this expert knowledge was added by highlighting remedies that the experienced homeopath was already considering intuitively. This intuition can be incorporated in computer-programs when analysed properly.(43)
The remedies in the repertory-rubrics have 3 (or 4) different degrees. The precise meaning of these degrees is ill defined. It has something to do with the common notions sensitivity and specificity. When we have validated all remedies it must be possible to replace these degrees by sensitivity and specificity.
The repertory is so voluminous that an alphabetical order of symptoms is not sufficient. There is also listing by body functions and body-scheme. To use the repertory properly one has to make much effort in learning how to handle it.

5.5. Ponderation factors

There are important and less important symptoms. There are remedies with many symptoms and remedies with few symptoms. A simple spreadsheet is not sufficient to handle these differences. Important symptoms and small remedies must be made more conspicuous. This is translated in different 'ponderation factors' and expert-systems in the computer-programs. For commercial reasons these instruments are offered as black boxes, giving the user no insight in the throughput process. For a better development of the method it is necessary that the algorithms are made public. We give some indications how these instruments can work

5.5.1. Small rubrics

Important symptoms are, according to § 153 of the Organon, special symptoms. Those correspond to small rubrics in the repertory. So small rubrics have to be made more conspicuous. According to mathematical information theory the information value of a phenomenon is higher if the phenomenon is more rare. The formula for this is: H(x) = - ld(p(x)): The information-value of symptom x is the inverse of the 2-logarithm of the chance for x. This means e.g. that the information-value of a rubric with 2 symptoms is 3 times the information value of a rubric with 8 symptoms.

5.5.2. Small remedies

We cannot say that a small remedy is less useful, nor that a small remedy is as useful as a polychrest. There are many cases solved by small remedies after failure of polychrests, but we cannot be conclusive about the value of those remedies. Small remedies are relatively more used in 'clinical' homeopathy. In general we experience that small remedies are lost in the process of repertorisation. So we need correction for the size of the remedy. This type of correction is not as straightforward as for small rubrics. There is a wide variety in the number of symptoms of each remedy; sulfur has nearly 9,000 symptoms and there are remedies with some 10 symptoms. In the original repertory of Kent 45% of all remedies have less than 100 symptoms. The symptoms are divided over 55 chapters in the repertory, but not evenly. Small remedies have concentrations of symptoms in certain chapters, especially when they originate from 'clinical' homeopathy. We want to apply weighing in order not to miss a remedy that has proven to be useful for diseases in certain organs. This can be done by detecting the number of symptoms of the remedy for the organ related to the total number of symptoms of the remedy. A similar problem concerns the symptoms in the chapter 'Mind'. Many remedies have few to none symptoms in this chapter. The remedies tend to get lost when we use predominantly mind-symptoms when repertorising.

5.6. Hierarchy of symptoms

The choice of remedies in homeopathy is very different from the choice of conventional medicines. The complaint or diagnoses plays a minor role in this. First of all we make an extensive description of the patient, including medical history, body-functions, mental and psychological functions and reactions to all kinds of environmental influences. Then we value the different features of the patient. Hahnemann taught us to value the peculiar symptoms the most.(44) Later Kent added another hierarchy: first mental symptoms, then general symptoms and at last local symptoms. However, when we analyse the work of experienced homeopaths we see some deviations from Kents' rules.(45)

5.6.1. Emotional and mental symptoms

These symptoms are indeed more important than local symptoms but their use depends on the phase in the diagnostic process. When we are putting up a differential diagnosis of possible remedies mental symptoms give more semantical problems. The symptom 'Irritability' has many synonyms in different rubrics (with different remedies). After the differential diagnosis has been set however we use irritability to confine our choice and prefer the remedies where irritability is an important characteristic.

5.6.2. Essence

Many remedies have an 'essence'. This is a combination of all sorts of symptoms characteristic for a certain remedy. The symptoms in this essence can be local and therefore have a higher value as when we consider this local symptom apart from the other symptoms.

5.6.3. Local symptoms

As seen in the former paragraph local symptoms have a higher value when constituting the essence. But there is another reason why local symptoms must be upgraded. Sometimes several local symptoms appear in a fixed combination related to a certain remedy. Like the combination of hoarseness and skin eruptions in the remedy Graphites. During the treatment a local symptom can gain importance when that symptom persists while other symptoms vanish.

5.7. Linguistics

Using communication and observation we try to make reproducible interventions. Most of it is based on communication. In the therapeutic process firstly the patient communicates his complaints and distinguishing marks. The doctor interprets this with his own background. At last a remedy is searched that seems to suit this interpretation in a text that has been compiled by others. So we have three phases that can by biased by natural language. Bias by natural language has three aspects:(46)

1. Syntax
2. Semantics
3. Context, words can get different meanings in different contexts

So communication is one of the most important issues in homeopathic science, relevant in two aspects:

1. Communication between doctors and between doctors and patients.
2. Efficiency of the method.

Communication between doctors, especially between homeopathic and conventional doctors, needs to improve to establish the proper place for homeopathy. But also for the development of the method language is crucial. We need to be more aware of the different aspects of natural language to improve all three phases of the homeopathic process: case taking, processing of symptoms an validating medicines (=creating materia medica).

5.7.1. Materia medica

With the computer we can search every word in the materia medica. At the moment this is scarcely used in daily practice due to insufficient precision and incomplete searches. When we study the materia medica we see many unnecessary synonyms, like 'urination' and 'micturition'. A thesaurus can be a solution for this problem, but not for homonyms. By correcting synonyms and homonyms in the materia medica we can make our search more precise and complete.

5.7.2. Repertory

After linguistic revision and validation of the materia medica we need a new repertory, assuming that information technology does not make any form of index redundant.
Before we depicted some problems regarding the repertory. The reason that we still use an outdated instrument so frequently is the additional value of the learning process to handle the repertory. The repertory is not just an index but stands for a large amount of associative knowledge structured by a book. We need thorough analysis of this associative knowledge before we can make a new repertory. Probably we need an expert-system instead of an index. But still we must expect that a new repertory will need a new learning process to handle it.

5.8. Registration

In Holland we have several software-programs to register casuistry, some commercially developed some by homeopathic physicians. On one hand we need standardisation to be able to exchange data. On the other hand these programs must provide efficient practice-management, so it has to suit the local situation. Both needs can be met but need close monitoring. Another concern is the financial aspects. For a software-developer it is difficult to develop profitably software, flexible to local needs, for a group of several hundred professionals.

5.8.1. Gathering

We have some experience with registration in homeopathic practice. To bring these data together is another story.(47) It's not a question of just adding up. This is not allowed for reasons of privacy. What we gather depends on our aims, for instance:

• Validating materia medica we need all cases scoring +3 or +4 (Dutch consensus) with a follow-up exceeding one year.
• To assess results by diagnosis we gather all casuistry per diagnosis considering inclusion and exclusion criteria.

Whatever aim we have, our demands will change in the process. This will have repercussions for the software. Will this be financially attainable. This pleads for careful analysis and not too elaborate software.
We need a powerful international institution to monitor registration and the tools for this purpose.

5.8.2. Interpretation

It's no use gathering data without interpretation. We have some experience in the use of parameters to evaluate the effectiveness of treatment.(48, 49) These parameters can be described well enough using disease classification. It is important to strive for a standard in classification.
Validating materia medica with the recorded data is a new development. This is more difficult because the goals are less concrete. One of the goals is the determination of confirming symptoms to increase the certainty of our prescription. Such a symptom might be 'Salivation in sleep' for Mercurius solubilis. If this symptom has a high sensitivity as well as a high specificity we can be pretty sure of the remedy when the symptom is present or discard the remedy when it is not present (see below). There are pitfalls here, e.g. we do not know what is not recorded. When 10 colleagues sit together to evaluate 20 cases of Mercurius solubilis it is possible that some colleagues did not ask for this symptom. So the symptom may be present in more cases but is not recorded. It is also possible that a symptom is incorrectly recorded after suggestive questioning. Such problems must be kept in mind during consensus meetings.

Another problem is confirmation bias:

• If 'Being precise' is the most important indication to prescribe Silicea, possibly Silicea is not prescribed to persons who are not precise. Maybe this is correct, maybe not. It seems interesting to search for Silicea patients who are not precise.
• Precise persons who did not respond well to Silicea are not presented during the consensus meetings. At the moment we only regard the successful cases.

As stated before, more knowledge leads to new problems. We must remain critical.

5.8.3. Specificity, sensitivity and likelihoodratio

Questions are our main instruments. We have different stages in our consultations: partly the questions are standardised, partly the questions are generated by our presumptions. As in conventional diagnostic instruments we can use questions to enclose or discard a remedy in our differential diagnosis. For that purpose questions have to be validated, e.g. by assessing symptoms in successful cases. This leads to the following statistical concepts:
Sensitivity of a symptom of a remedy is the percentage of people that need just that remedy and also have that symptom.
Specificity of a symptom is the percentage of people that do not have the symptom and do not need the remedy.

When we have validated a certain number of cases of a certain remedy we can calculate the sensitivity of the symptoms belonging to that remedy. Specificity can only be calculated if we have validated all remedies. In daily use however these concepts are a bit awkward. There is confusion about the exact meaning of these concepts. There is a more practical concept called likelihood ratio (LR). There is the likelihood ratio(+) to include a remedy and the likelihoodratio(-) to exclude a remedy. If we take a closer look at the constituting parameters in the likelihood ratio we see:
Likelihood ratio(+) = (prevalence of a symptom with the remedy) / (prevalence of the symptom in the rest of the population). This means: If the symptom is more frequently present with the remedy than with the rest of the population the LR(+)>1. In other words the more the symptom is confined to the remedy (and not to the rest) the higher the likelihood ratio(+). In fact the likelihood ratio(+) is a mathematical representation of §153 of the Organon by Samuel Hahnemann that states: pay particular attention to the peculiar and characteristic symptoms(50).

Likelihood ratio is a better measure for the importance of a remedy in view of a certain symptom than the typeface that is used in our repertories. The importance of a symptom is now based on opinion. In the long run we aim at a repertory based on experimentally validated data. This repertory can suggest to us questions to ask about certain remedies to enclose or to discard that remedy.

5.9. Homeopathic expertise

We apply statistics to make our instruments, materia medica and repertory, more accurate. Another goal of our consensus meetings is to evoke meta-knowledge of remedies by increasing awareness and recognition of patterns. The development of homeopathic expertise is the same as the development of conventional medical expertise: In the first phase fragments of medical knowledge are stored in memory as separate facts. By accumulation of a number of relevant facts it becomes possible to reason out of these facts and some logical structure unfolds, a so-called network of knowledge, where knowledge gets a logical position. This is the second phase in learning. Our materia medica and repertory are examples.

In the third phase pattern-recognition comes forward: Frequent use of networks of knowledge gives rise to condensed conglomerates of knowledge: reasoning is not necessary because patterns are recognised at first glance, meaning that disease or that remedy. The pattern of a disease and its manifestations is called illness-script. The experienced GP recognises the appendicitis when the patient is still in the waiting room.

The memory of experienced physicians gets more and more occupied with such illness-scripts. In the forth and last phase the classical knowledge of the books figures only in the background. Illness-scripts have become more complex and variable by added experiences making the physician capable of recognising the pattern even in complex and exceptional situations. There is a great gap between this kind of knowledge and the information in the books; the elements in the pattern and the starting-points are hardly traceable.

All this can be compared with the way experienced bird-watchers make correct observations. In a flash the bird is recognised. His clues are in the appearance of the bird, the flight and the sounds, but also the context of the observations are taken into account. This is performed in split seconds.

5.9.1. Patterns of remedies

In homeopathy most physicians are in the second or third phase of knowledge. New knowledge about existing remedies or (still) unknown remedies is stored in the brain in a somewhat encyclopædial way. Recognition will only take place after ample experience with the remedy. Some remedies are seldom used, once a year by one doctor or less. Other remedies are used more often, but a pattern is not yet established. By bringing doctors together we hope to enhance the recognition of patterns.

5.9.2. Misguiding expertise

Following family medicine homeopathy becomes aware of its own scientific identity, discovering necessary elements in other fields of science. We hope that our colleagues can agree about the need for a proper scientific identity for homeopathy. If so a scientific plan of action must be composed. Such a plan must have the support of the profession. In this paper we merely want to communicate our views according to our experience. Some recommendations are made in the last chapter. We want to make some final remarks on function and organisation of a scientific policy.

6.1. Justification

An important issue in scientific approach is justification; do we do the right things? We have to render account for our doings towards patients, colleagues and financiers. For those different groups accountability means different things:
Patients: what can a patient expect from homeopathy and/or its different views? Can it be applied in every disease? Also: What is the meaning of 'measurements', is it real exact measuring or commuting?
Colleagues: presentation of new ideas must be accompanied by the invitation to falsify these ideas. This falsification must be made possible by clear descriptions of theories and method so that anyone can verify it (do we do the right things or do we do things right?). The presentation of a large number of new remedies at the same time must be avoided because this discourages falsification. Anyone must be able to learn a 'measuring'-method that pretends exactness.
Insurance companies: this group has the right to know about effectiveness of homeopathy.

6.2. Recognition

Recognition of non-conventional methods does not solve the problem of non-conventional medicine. Within a non-conventional method there are always even less-conventional points of view. Recognition of a method only causes a shift of the border between conventional and non-conventional. The struggle for recognition appears more like a struggle for power than a scientific dispute. Recognition leads to a substantial bonus expressed in money and infrastructure. These facilities are necessary for good research. This also has a circular effect.
Considering this it seems better to create a step-by-step recognition-ladder. Each step of this ladder entitles to more facilities in infrastructure and funding.

6.3. Finance

Regarding funding for research there are essential differences between homeopathy and allopathy. There are also differences in homeopathic medicines for self-medication (complexes) and singular medicines. For this paper we consider singular remedies.

These points apply to allopathy:

• Medicines are developed for indications.
• Practitioners have a straightforward choice in medicines
• Medicine are protected by registration and patents.
• Dosage is frequent

These points apply to homeopathy:

• There are no specific indications. One cannot say beforehand if there is a market for a new medicine.
• Medicines are cheap so funds to spend for research are small.
• Dosage is infrequent.

While allopathic medicine have a daily turnover of tenths of Euros per patient, the daily turnover of a homeopathic medicine per patient is insignificant. Therefore we must consider other means of financing scientific research. Financing by insurers is one of the possibilities, also guaranteeing more objective research.(51) In the end the user of homeopathy pays for research, be it by government-taxes, by medicines or by insurance.

6.4. Organisation

This paper is not meant as a complete plan of action. There may be other issues of interest. From this paper we derive some recommendations:

1. We need to investigate if homeopathy has a different paradigm.
2. Homeopathy deserves a proper scientific identity and therefore a scientific institute.
3. The method is far from perfect and needs improvement with suitable tools.
4. Proving the efficacy of the method is at the moment less important than improving the efficiency.
5. The responsibility of each homeopathic physician regarding the development of the method must be clear from the start, i.e. from the education onwards.
6. The quality and the description of each case is the cornerstone of our scientific development.
7. The use of hypotheses must be restricted.
8. We need a priority-list of hypotheses to be validated.
9. All casuistry that is made public must be peer-reviewed.
10. Semantics and confirmation-bias require constant attention and exchange of knowledge with other disciplines.
11. Consensus procedures must be worked out and embedded in the profession.
12. We need a list of remedies to be validated. The necessity of the development of new remedies must be proved.
13. Validation is a circular process and needs regular evaluation.
14. The concept ‘important symptom (§153)’ must be replaced by likelihood ratio(+).
15. The workload must be divided internationally.
16. The materia medica and the repertory must be rebuilt.
17. We need databases for casuistry and for literature.
18. Art and science must be regarded as complementary.

8. Literature

1. Kleijnen J. What research is needed to show the effectiveness of homeopathy? BHJ 2000;89(suppl.1):S1-2
2. Schouten JSAG, Meer J van der: 'Volksgezondheid toekomst verkenning ' 1997. VI. Effecten van zorg. NTvG 1998/23;1335-8
3. Wiersma T. Twee eeuwen zoeken naar medische bewijsvoering. Amsterdam: Boom/Belvedère, 1999
4. Brands M. Disease: language and experience. A cognitive comparison of allopathy, homeopathy and Chinise medicine. Thesis, Amsterdam, 1998
5. Kuhn TS: The Structure of Scientific Revolutions. Chicago; University of Chicago Press, 1962. Chapter 6
6. Boers M. Dwalingen in de methodologie. XXV. Uitkomstmaten, surrogaateindpunten en intermediaire maten. NTvG 2000;41:1949-52
7. Kent JT. Lectures on homeopathic philosophy. Thorsons Publishers Wellinborough. 1979. Hoofdstuk XXXV.
8. Senge PM. De vijfde discipline. De kunst & praktijk van de lerende organisatie.1992 Scriptum Management ISBN 9071542 54 8
9. http://skepdic.com/confirmbias.html
10. Gilovich T. How We Know What Isn't' So: The Fallability of Human Reason in everyday Life. New York. The Free press, 1993
11. Sadler JZ. Recognizing values: a descriptive-causal method for scientific discourses. The Journal of Medicine and Philosophy 1997;22:541-65
12. Stolper E Kipp R Lugten R. Materia Medica Validering: Silicea en Conium. SSC 1999/2; 11-12
13. Hahnemann S. Organon der Heilkunst. Heidelberg: Haug. 1974
14. Gaarder J. De wereld van Sofie. Houtekiet/Fontein 1997
15. Grindmeijer Reenders Rutten. Het geneeskundig proces; van klacht naar therapie. 1999
16. Wiersma T. Homeopathie als verboden spiegelbeeld van de reguliere geneeskunde. Kennis en Methode 1988/4, 295-314
17. Gleik J. Chaos. London: Sphere books, 1998
18. Lo S-Y, Lo A, Chong LW, Tianzhang L, Hua LH, Geng X. Physical properties of water with IE structures. Modern Physics Letters B; 1996/19: 921-30
19. Demangeat JL, Demangeat C, Gries P, Poitevin B, Constantinesco A. Modification des temps de relaxation RMN a 4 Mhz des protons du solvent dans les tres hautes dilutions salines de silice/lactose. J Med Nucl Biophy 1992; 16: 135-45
20. Endler P Schulte J, eds. Ultra high dilution physiology and physics. Dordrecht: Kluwer Academic, 1994
21. Stolper CF Kipp R en Lugten RGF. Consensus 'Casus geslaagd of niet'. SSC 1997 27/1, 21-22
22. Lugten RFG Kipp R en Stolper CF. Wie geneest heeft gelijk. SSC 1997 27/1, 23-25
23. Craen AJM de. Dwalingen in de methodologie. XVII. Het placebo-effect. NTvG 2000;7:322-4
24. Guajardo e.a. Homeopathic terminology: a consensus quest. British Homeopathic Journal 1999/3, 135-41
25. Jansen Gerard, Biolchini Jorge en Stolper Erik. Van Registratie naar (onder meer) Indicatie. SSC 1999 29/1, 2-5
26. Riley DS. Homöopathische Arzneimittelprüfungen - Grundlagen und Praxis. Homint R&D Newsletter 1/96, 3-14
27. Koster D Haselen R Jansen G Dicke M.Homeopathic medicine test according to the crossover design: from theory to pratice. Br. Hom. J. 1998/87; 181-9
28. Frei H Thurneysen A. Homeopathy in acute otitis media in children; treatment effect or spontaneous resolution? British Homeopathic Journal 2001;90:180-2
29. Frei H Thurneysen A. Treatment for hyperactive children: homeopathy and methylphenidate compared in family setting. British Homeopathic Journal 2001;90:183-8
30. Wijk R, Wiegant FAC. The similia principle in surviving stress; mammalian cells in homeopathy research. Utrecht University 1997
31. Eisebitt R. Methods used in studies of health economics in complementary medicine, with special reference to homeopathy. Homint R&D Newsletter 1/99, 4-11
32. Dantas F. Reporting and investigating adverse effets of homeopathy. British Homeopathic Journal 1999/3, 99-100
33. Stolper CF. De kleren van de keizer. Een kritische beschouwing. SSC 1994/24,, 12-18
34. Rutten Lex. De helft van wat ik zeg … SSC 1996;2:12-5
35. Cure N. Trials and tribulations of the homeopathic materia medica. British Hom. Journal 1989: 156-65
36. Scholten Jan. Homeopathie en mineralen. Utrecht1993
37. Scholten Jan. Homeopathie en de elementen. Utrecht 1996
38. Borneman JP. Blackie Memorial Lecture 2001. Homeopathy: progress and promise, a critical perspective. British Homeopathic Journal 2001;90:204-13
39. Rutten Lex Macht's nach, aber … SSC 2000;30:4-7
40. Several authors. Mittelverwechslungen im Reprtorium von Kent. Zeitschrift fúr Klassische Homeopathie; 1989, 1990, 1991
41. Rutten Lex. Wie is bang voor het repertorium? SSC 1987/17, 43-9
42. Rutten Lex. Wie is bang voor het repertorium? (2) SSC 1987/17, 116-23
43. Rutten Lex. Samuel: a new software package. Jrn. of the American Inst. of Homeopathy 1986/79, 33-9
44. Hahnemann S. Organon der Heilkunst. § 153. Heidelberg: Haug. 1974
45. Rutten Lex. Hidden paradigms in homeopathy. British Homeopathic Journal 1991/80, 39-41
46. Allen J. Natural language understanding. The Benjamin/Cummings Publ. Comp. Menlo Park, California. 1987
47. Rutten Lex. Hoeveel tijd kost registreren. SSC 1999 28/1, 6-9
48. Berckel Smit JACM van. Apilot study evaluating the efficacy of homeopathy in daily practice. Br. Hom. J.1993 82/1, 9-15
49. Fruijtier PC Berckel Smit JACM van en Brink EA. Effectiviteitsonderzoek in de homeopathische praktijk. SSC 1998 28/2, 11-13
50. Stolper CF, Rutten ALB, Lugten RFG, Barthels RJWM. Improving homeopathic precribing by applying epidemiological techniques: the role of likelihood ratio. Homeopathy;91:230-238
51. Cate JW ten. Klinisch onderzoekers en farmaceutische industrie. NTvG 1999; 143/26, 1349-52

Lex Rutten, MD

Aard 10 - 4813 NN Breda, Netherlands