Complementary Therapies in Medicine (2013) 21, 565—570

Available online at www.sciencedirect.com

ScienceDirect journal homepage: www.elsevierhealth.com/journals/ctim

COMMENTARY

The importance of case histories for accepting and improving homeopathy Lex (A.L.B.) Rutten ∗ Committee for Methods and Validation of the Dutch Homeopathic Doctors’ Association, Aard 10, 4813 NN Breda, The Netherlands Available online 17 October 2013

KEYWORDS Homeopathy; Case history; Bayes’ theorem

Summary Case histories are necessary besides other types of evidence to convince doctors of a specific type action of homeopathic medicines. Prognosis of treatment does not merely depend on efficacy. Some considerations based on consensus meetings about best cases and prospective research into the relationship between symptoms and result. Many data in homeopathic literature are unreliable because of wrong interpretation, insufficient numbers and confirmation bias. Causal relationship between medicine and ‘cure’ could be documented better. Extraordinary cases are not helpful to increase reproducibility. Conclusion: For acceptance and improvement of homeopathy cases should be reproducible. ‘Normal’ cases reflecting daily practice contribute more to this goal than extraordinary cases. Accuracy can be increased by larger samples of comparable cases. Causal relationship between medicine and improvement should be further explored. © 2013 Elsevier Ltd. All rights reserved.

Introduction Several renowned epidemiologists stated that the proof for homeopathy is not inferior to the proof for conventional medicine.1,2 Others state that homeopathy is a placebo effect.3—5 Clearly, there is subjectivity involved in the interpretation of scientific evidence.6 A major problem is the plausibility of homeopathy’s mechanism of action.7 Homeopathic doctors, however, experience that the effect of homeopathic medicines is different from conventional medicines and sometimes very unexpected. Homeopathic



Tel.: +31 765227340. E-mail address: [email protected] URL: http://www.dokterrutten.nl/Inc.html.

physicians acknowledge that potentised medicines cannot (and do not) work like conventional medicines. Case histories show the benefits of homeopathy to doctors without homeopathic experience, but they also show the ingredients of successful prescriptions to experienced homeopathic practitioners. Case reports are still important in conventional medicine too, especially for discovering the unexpected.8 Both case histories and RCTs have their limitations, but also their advantages, insufficient understanding about this leads to subjective interpretations of results. An important limitation of RCT is the fact that an RCT is confined to a specific condition and co-morbidity is a reason for exclusion, while co-morbidity (multi-morbidity) is an important reason to choose homeopathy. Recently, prognosis research has become a priority in clinical research and practice.9 Prognosis research aims ‘‘. . .to help improve the evidence base

0965-2299/$ — see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ctim.2013.10.001

566 for the information given to patients about their disease and guide clinical decisions about treatment’’. This is an interesting development for case histories illustrating successful cases. To be meaningful case histories must pay attention to some points: - Indications of the causal relationship between medicine and improvement - the importance for daily practice - concordance with similar cases - how homeopathy works in daily practice. But the impact of case histories increases if colleagues can reproduce these results. To be really valuable in this respect, case histories need some extra requirements: - a clear algorithm; what information led to the choice of the medicine and why? - reproducibility: if the reader has a similar case he will probably have a similar result - valid information. Increasing reproducibility could be the most compelling mission for homeopathy’s future. The basic information about homeopathic medicines should be valid, easy to learn and readily accessible. The validity of information from case histories is not self-evident. For instance: does a cure of headache in one case mean that the prescribed medicine cures headache? A systematic error in the homeopathic database is the fact that entries of repertory-rubrics are hitherto based on absolute occurrence in successful cases. This way a frequently occurring symptom will be added to the materia medica of every medicine in the long run and a considerable part of the information in homeopathic repertories becomes misleading.10 This way it makes no difference if a symptom is seen in one out of hundred or in one out of three cases. Bayes theorem states that entries should be based on relative occurrence. Another systematic error in homeopathic data is confirmation bias: observations are influenced by existing ideas and experience. Besides systematic error (bias) there is statistical variance. This is an important source of invalid information in homeopathic data, because much information is based on a limited number of cases. Many practitioners like to see extraordinary cases, but everyday cases might be more valuable to the improvement of homeopathy. Case histories become more valuable when similar cases are brought together to enlarge a specific population. Similarity can be found in the same medicine, the same condition, but also in the symptoms characterising the case. This paper aims to show how we can increase the reproducibility of homeopathy by case histories. This reproducibility depends, among others, on a Bayesian algorithm to handle decisions based on multiple variables, as the choice of a homeopathic medicine is based on a combination of symptoms and characteristics that indicate a specific medicine. If we can extract the prognostic factor of symptoms we use in homeopathy we can make predictions about the chance a medicine will help.

L.(A.L.B.) Rutten

Materials The position of case histories is illustrated by two projects of the Committee for Methods and Validation of the Dutch association of homeopathic physicians to validate homeopathic data: 1. Retrospective case analysis: consensus meetings to evaluate best cases of specific medicines (MMV project). In the MMV project Dutch doctors were invited to present their best cases concerning homeopathic medicines. Best cases concerning some 25 medicines were evaluated by peers. 2. Prospective assessment of the relationship between six homeopathic symptoms and treatment outcome (LRproject): an observational study conducted from June 2004 until December 2007 including all consecutive new patients older than two years. The goal was assessing the relationship between symptoms and successful outcome. Observers were 10 Dutch medically qualified doctors with more than 10 years experience in homeopathy and already participating in the MMV project. Six symptoms were recorded: ‘Diarrhoea from anticipation’, ‘Fear of death’, ‘Recurrent herpes of the lips’, ‘Grinding teeth during sleep’, ‘Sensitivity to injustice’ and ‘Loquacity’. At the end of the LR-project 4072 prescriptions concerning 4094 patients were evaluated. Discussions in the first project about how we apply our experience in daily practice resulted in discovering an algorithm expressed by Bayes’ theorem. The discussions went on during the prospective research and we discovered pitfalls in the way we observe cases and the way experience is entered in our instruments, Materia Medica and Repertory.

Improving homeopathy The projects described above were not meant to prove homeopathy, but to improve the method. We tried to analyse the meaning of experience and to become more aware of bias in our observations and conclusions drawn from these observations. Discussing cases in the first project the question ‘was the improvement in the case really due to the prescribed medicine?’ always came first. These discussions, also guided by literature studied by the group, made clear that we fool ourselves if we ‘polish’ cases to ‘prove’ a certain point or to make the case more impressive. Only ‘real’ information can be reproduced. Consensus meetings were also held during the prospective research, based on reports on interim results, discussing differences between participants. It is, e.g. amazing how much the interpretation of one symptom can vary. There was also no intention to treat analysis of data; only cases we considered good enough were used to calculate results. Our aim was to collect as many data as possible without interfering with daily practice, realising that there is a conflict between validity and feasibility.11 The main purpose of our research was to discover how symptoms are related to successful cases.

The importance of case histories for accepting and improving homeopathy

Prognosis research Future outcome of a disease depends on several variables, efficacy of a medicine is only one of them. Co-morbidity, adverse effects and genetic susceptibility may influence the effectiveness of a medicine. Prognosis research is especially interested in clinical outcome. Knowledge about homeopathic medicine is partly based on successful cases, good clinical outcome. The Materia Medica of a homeopathic medicine comprises symptoms that indicate successful outcome. Research into variables that influence outcome is called prognostic factor research.12 A homeopathic symptom is a prognostic factor. The probability that a specific medicine will work increases step by step if the patient reveals new symptoms that indicate that medicine. How this works is clarified by Bayes’ theorem.

Bayes’ theorem In our consensus meetings we asked the participating doctors why a symptom is an indication for a specific homeopathic medicine. Doctors answered something like: ‘‘In my experience patients that respond well to this medicine have this symptom more frequently than other patients’’. In other word: the prevalence of the symptom is higher in the population that responds well to the medicine than in the remainder of the population. In statistics this is expressed as the Likelihood Ratio (LR) of the symptom. So, experience with a specific medicine leads to recognising new patients that will probably respond well to that medicine. This process is similar to the medical diagnostic process: if abdominal motion pain occurs frequently in patients with appendicitis, the doctor will suspect appendicitis in the next patient with motion pain. 13 How we learn from experience can be explained by Bayesian philosophy. Bayes’ formula is derived from the statistical formula of conditional probability. It tells us how experience can be translated into future expectations. In Bayesian statistics new facts or observations update probabilities, with the following formula1 : Posterior odds = LR × prior odds In words: the probability of a fact increases if it’s prevalence in a test or observation is more than expected from prior knowledge. Bayesian statistics provide an algorithm for learning from experience. Bayesian statistics also offer an algorithm for homeopathic medicine selection. An example: Prospective research showed that 14 out of 42 patients (33.3%) responding well to the homeopathic medicine Lachesis were loquacious, while this number was 253 out of 4052 (6.2%) for the remainder of the population. LR = 5.3 (95% Confidence Interval 3.4—8.3). With this knowledge you can adjust the expectation that Lachesis will work if the new patient

1 Odds = chance/(1 − chance); in words: the chance that something will happen divided by the chance that it will not happen. Odds = 1 means: chance is fifty—fifty. Chance = odds/(1 + odds). LR = Likelihood ratio = true positives/false positives.

567

before you is loquacious. If, say, your prior expectation that Lachesis will work was 5% (odds = 5/95) your posterior expectation with the symptom loquacity becomes 21.8% (posterior odds = LR × prior odds = {5.3 × 5)/95} = 0.279. Then Posterior chance = 0.279/(1 + 0.279) = 21.8%. Now suppose the patient has two other symptoms characteristic for Lachesis: mainly left-sided complaints, estimated LR = 4, and intolerance to tight clothing, estimated LR = 5. After each new symptom the chance that Lachesis will work is updated. For this patient the estimated chance that Lachesis will work develops with three consecutive symptoms as described in Table 1. It is generally accepted in homeopathic practice that after three symptoms characteristic for a specific medicine the doctor becomes rather confident that the medicine will work. Homeopathic symptoms are in fact prognostic factors; instead of diagnosis they indicate prognosis. These calculations show that homeopathy obeys to a sound algorithm and can easily be reproduced if LRs are valid. At the moment most homeopathic data are estimates, but they can be scientifically assessed. More accurate data will increase the reproducibility and effectiveness of homeopathy. Accuracy increases with larger samples and avoiding bias.

Bias in homeopathic data The homeopathic Materia Medica consists of symptoms and personal characteristics that indicate a specific medicine. Such symptoms and characteristics occur more frequently in patients responding well to that specific medicine than in the remainder of the population. This principle of relative occurrence has a sound theoretical basis in Bayes’ theorem explained above. However, homeopathic experience is hitherto recorded as absolute occurrence not related to the prevalence of the symptom in the general population. Homeopathic repertories indicate that the symptom ‘fear of death’ can occur in patients responding well to medicine X. But they should indicate if this symptom occurs more frequently in the population responding well to medicine X than in the remainder of the population. In the present situation especially frequently used medicines are increasingly mentioned falsely in relation to frequently occurring symptoms. This situation hampers the efficacy of homeopathy, but is especially confusing for novices in the method.

Chance and homeopathic data Now we have an algorithm for learning from experience: the prevalence of a symptom in the target population should exceed the prevalence in the remainder of the population. But how can we know the prevalence of a symptom? During our consensus meeting participating doctors gave estimates of the prevalence of symptoms. There was fair consensus in these estimates: the prevalence of ‘keynote-symptoms’ (characteristic symptoms) in homeopathy was mostly estimated to vary around 5—10%. This is understandable from a Bayesian point of view: such symptoms are seen in every 10—20 patients and LR values could theoretically reach 20 if the symptom occurs in all patients in the target group and in 5% of the general population. The estimated prevalence

568

L.(A.L.B.) Rutten

Table 1 Algorithm for homeopathic medicine selection. The chance that a specific medicine will work increases with each symptom that is characteristic for this medicine. Symptom nr.

Prior chance (%) *

1 2 3

5 21.8 52.7 *

Symptom

LR

Posterior chance (%)

Loquacity Left-sided complaints Intolerance to clothing

5.3 4* 5*

21.8 52.7 84.7

These are estimates.

of some symptoms in the general population was confirmed in our prospective research. But how can we estimate the prevalence of a symptom in the population responding well to medicine X? This is impossible in one case, but even if you bring cases together we must be careful. An example: During a Dutch consensus (MMV project) meeting 15 doctors presented 23 good Sulphur cases. Only one of the 23 patients (4.3%) had the symptom ‘fear of death’, and the group therefore concluded that ‘fear of death is not an important characteristic for Sulphur. But the doctor who presented this case had two Sulphur cases. For this doctor ‘fear of death’ was present in 50% of the Sulphur cases, and therefore until this meeting an important indication for Sulphur. In prospective research assessing the symptom ‘fear of death’ out of 88 Sulphur cases only one (1.1%) had the symptom ‘fear of death’. The prevalence of ‘fear of death’ in the whole population of 4094 patients was 3.9%, LR = 0.29 (95% CI 0.041—2.048). The symptom ‘fear of death’ is therefore a contra-indication for Sulphur. This example shows how individual doctors can be misled by personal experience and how increasing numbers by aggregating such information can resolve this problem: we need many comparable cases and systematic assessment of specific symptoms. The difficulty of collecting enough cases must not be underestimated. In the Dutch LR project 10 experienced doctors evaluated 4074 prescriptions over three and a half years. In total 421 different medicines were used and only 75 medicines accounted for more than four successful cases.

Causal relationship During our consensus meetings each case was assessed by peers and many cases were not accepted because the improvement could have been caused by other factors. Such assessments are not feasible in large data collections and therefore each doctor should realise his responsibility to make an honest assessment of causal relationship in each case. In our prospective LR assessment all participants were trained in this respect during the preceding MMV consensus meetings and the data were collected twice a year and discussed with the research participants. The participants were well instructed about the negative influence of wrongfully allocated cases.14 All participants agreed about some criteria that indicate causal relationship: improvement resumed after repeated administration of the medicine, the effect comprised more than the presented complaint, and the course of improvement followed ‘Hering’s rule’. Hering’s rule describes developments that indicate good prognosis as a result of a homeopathic medicine. The

most valued development is improvement from inside outward: in a case with asthma and eczema the asthma should improve first. This fact, and the fact that the same medicines improve both asthma and eczema, is regarded to indicate an effect of a homeopathic medicine. Such criteria are also used in assessing causality in adverse effects of medicines, see later in the discussion section.

Confirmation bias Some symptoms in the homeopathic Materia Medica are so closely associated with a specific medicine that the medicine is not considered if that symptom is absent. An example is the symptom ‘sensitivity to injustice’ for the medicine Causticum. In our prospective assessment of the relationship between symptoms and success the symptom, ‘Sensitivity to injustice’ was present in every patient prescribed Causticum at the onset of the study, but after two years the prevalence of the symptom was only 40% in the population responding well to Causticum. We saw similar results in our MMV consensus project: in the 10 best cases of Causticum with more than one year follow-up ‘sensitivity to injustice’ occurred in 40% (4/10). Confirmation bias could restrain us from prescribing the right medicine if an expected symptom is missing. But it will also lead to misleading data in case histories and therefore to less reproducibility. In our experience confirmation bias can be prevented by demanding longer follow-up for case histories. Longer follow-up will also help to establish the causal relationship between improvement and medicine.

Discussion Although the RCT evidence for homeopathy is not inferior to evidence for conventional medicine this evidence does not convince. To feed the open mind, or maybe to open up the mind, we need case histories that show how homeopathy differs from conventional pharmacology and how homeopathic medicines should be prescribed. Case histories must tempt and enable others to reproduce similar results. On the other hand, homeopathic doctors must be prepared to think more scientifically. Many ‘cures’ are not their work. They should acknowledge the possibility of bias, like confirmation bias, and the misleading aspects of chance variance. They should also think more strategically and keep a clear purpose of case histories in mind. An interesting case might be contra-productive if nobody can reproduce a similar case. If your success was based on your unique skill, your intuition, or maybe mere chance, you may earn the deepest respect for your personal skills, but at the same time cause

The importance of case histories for accepting and improving homeopathy a hopeless feeling among colleagues like ‘‘This is impossible for me to reproduce’’. Then homeopathy becomes a sort of magic not suited for daily practice. It may seem boring to mention that the patient has a desire for salt in each Natrium muriaticum case, but it is very relevant to know if 30% or 60% of the Natrium muriaticum cases have this symptom. If we include this knowledge in our algorithm we can show interested colleagues how this symptom contributes to the choice of this medicine. Even for experienced homeopaths this knowledge is useful and hitherto unavailable. The requirements of reproducibility and statistical validity have several consequences for the content of case histories and for case taking in order to get suitable case histories. For the reproducibility of the case we should have two minimum requirements: 1. The ‘trigger symptoms’: what symptoms or symptomcombination made me think of this medicine? 2. Confirmatory symptoms: what symptoms that are well known for this medicine were present in this case?

569

enough to think of the medicine Natrium muriaticum? Or, when is this desire so low that we hesitate about the suitability of Natrium muriaticum? It may be tempting to describe a case as confirmed by many symptoms that are characteristic for the chosen medicine by lowering such cut-off values. This may help to make the picture of the medicine clear for the presentation of an individual case, but it will destroy the statistical reliability of our data. It is not easy to define exact cut-off values; how many salt should one take to call it a desire, and is this amount the same for children or in different cultures? In practice we intuitively deal with such questions by estimating the relative intensity of a symptom like: ‘‘This amount of desire occurs only in 5% of all comparable patients in my practice’’. Intuitively we understand, and this is confirmed by Bayesian theory, that a stronger symptom is a stronger indication for the related medicine.

Causal relationship

There are of course many other possible requirements to think of, like contradicting symptoms, but for the purpose of reproducibility these two should be noted in each patient journal. Apart from necessary content of the data we should also consider the quality of the data to get reproducible case histories. Quality of case descriptions could be improved by defining cut-off values for symptoms and causal relationship between medicine and effect. It might be better to make a strategic choice about what medicines to include in the database to provide a limited but successful set of medicines. Most important is a scientific attitude towards cases that are submitted.

Each case in our database that is in reality not improved by the homeopathic medicine is contra-productive, even if the story is very suggestive of a curative effect of the medicine. The supplied information in the description of the case is wrongfully allocated and could therefore be misleading. In our Dutch retrospective and prospective research projects the importance of causal relationship was well discussed and trained, but is this also possible in large international data-gathering projects? For assessing causal relationship in homeopathy it might be useful to develop algorithms. Algorithms for establishing causal relationship, like Naranjo’s algorithm, are already available for adverse drug reactions.15,16 Such algorithms could be adapted for homeopathy, an example is shown in Table 2.

Cut-off values

What medicines?

For most symptoms cut-off values are rather subjective. Many patients like salt, but when do we call it a desire strong

An experienced homeopathic doctor will not be amazed by a strong case of a well known homeopathic medicine;

Table 2

1. 2. 3. 4.

Proposed adaptation of Naranjo’s algorithm for homeopathy.

Was the case similar to other cases with this medicine? Did the effect appear after administration of the medicine? Did the effect after one dose subside after a period of time? Was the improvement resumed after repeated administration of the medicine? 5. Was there an initial aggravation? 6. Did the effect comprise more than the presented complaint, e.g. wellbeing and other complaints, like in the scale above 7. Did the course of improvement follow Hering’s rule? 8. Did old symptoms reappear for a while in the course of the improvement? 9. Are there alternate causes (other than the medicine) that solely could have caused the improvement? 10. Did the patient have the same response to other homeopathic medicines? 11. Was the effect confirmed by objective evidence?

Yes

No

Don’know

+1 +1 +1 +2

0 −1 0 −1

0 0 0 0

+1 +2

0 0

0 0

+2 +1

0 0

0 0

−3

+1

0

−1

+1

0

+1

0

0

570 he will be more interested in cases regarding unknown medicines. But what is the strategic value of cases of rarely used or even new medicines that are very hard to reproduce by inexperienced or beginning homeopathic doctors? Maybe it is better to get rid of misleading information about frequently used medicines first, before entering such misleading information about new medicines. In the Dutch LR-project only 50 medicines accounted for 72% of all successful prescriptions.17 Improving the use of these 50 medicines will have the largest effect on the effectiveness of the method. For strategic reasons it is better to create a separate database of cases for experienced homeopaths.

To prove or to improve? Seemingly contrary to what is said here about the importance of cases to convince others, it should be stressed that cases must not be described with the urge to convince non-believers in mind, but with the need to improve the homeopathic method. A case history may be very convincing on paper, but if the data are exaggerated it will be impossible for others to get similar results. This will work against accepting homeopathy. A reproducibility-strategy also implies that reproducible data must be included in the case history. It may be tempting to mention only the symptoms that make a case interesting — especially rare symptoms — but most prescriptions are based on ‘daily’ occurring symptoms. There seems to be some reluctance to share cases with colleagues. Maybe the need to improve the method is not clear enough. Maybe professionals think that only spectacular cases or cases about unknown medicines are relevant to submit. Several homeopathy journals tend to publish mainly extraordinary cases, but this policy should be reconsidered. What is the use of a case considering a medicine that you will prescribe only once a year against a case considering a medicine that we use every day? If we understand the basics of chance and statistics shown above we must realise that we need many cases concerning the same medicine. Other reasons why there is insufficient affluence of cases are practical. You need time to write down the case, you need to learn the skills and the format how to do this. Where should you send it to, and what happens with your material? Incentives must be convincing, and submitting cases should be as easy as possible to a database accessible for every practitioner. A case database should be carefully monitored and maintained, including feedback for doctors that submit cases. Missing data should be recovered. If there is a sufficient number of cases concerning one medicine prevalence of indicative symptoms should be calculated. Most of all: results should be published and be accessible. All these requirements demand a comprehensive logistic organisation.

Conclusion Case histories are important for the acceptation of homeopathy and for improving the method. From these perspectives reproducibility is the key-requisite for case histories. The data must represent the truth as good as

L.(A.L.B.) Rutten possible avoiding confirmation bias and cures not related to the medicine. The data should be complete regarding characteristic symptoms. Larger numbers diminish mistakes due to chance variance. In this respect strategic choices also imply confining us to a set of medicines that are frequently used. Bayes’ theorem provides an algorithm for interpretation of clustered cases.

Conflicts of interest None declared.

References 1. Kleijnen J, Knipschild P, Riet G ter. Clinical trials of homeopathy. Brit Med J 1991;302:316—23. 2. Vandenbroucke JP. Medical journals and the shaping of medical knowledge. Lancet 1998;352:2001—6. 3. Shang A, Huwiler-Müntener K, Nartey L, Jüni P, Dörig S, Sterne JAC, et al. Are the clinical effects of homeopathy placebo effects? Comparative study of placebo-controlled trials of homeopathy and allopathy. Lancet 2005;366: 726—32. 4. Altunc U, Pittler MH, Ernst E. Homeopathy for childhood and adolescence ailments: systematic review of randomised clinical trials. Mayo Clin Proc 2007;82(1):69—75. 5. Ernst E. Homeopathy: what does the ‘‘best’’ evidence tell us? Med J Aust 2010;192:458—60. 6. Lüdtke R, Rutten AL. The conclusions on the effectiveness of homeopathy highly depend on the set of analyzed trials. J Clin Epidemiol 2008;61:1197—204. 7. Vandenbroucke JP, Crean J.M. de. Alternative medicine: a ‘‘mirror image’’ for scientific reasoning in conventional medicine. Ann Intern Med 2001;135:507—13. 8. Vandenbroucke JP. In defense of case reports and case series. Ann Intern Med 2001;134:330—4. 9. Hemingway H, Croft P, Perel P, Hayden JA, Abrams K, Timmis A, et al. Prognosis research strategy (PROGRESS) 1: a framework for researching clinical outcomes. Brit Med J 2013; 345:e5595. 10. Rutten ALB, Stolper CF, Lugten RF, Barthels RJ. Statistical analysis of six repertory-rubrics after prospective assessment applying Bayes’ theorem. Homeopathy 2009;98:26—34. 11. Rutten ALB, Stolper CF. Diagnostic test evaluation by patientoutcome study in homeopathy: balancing of feasibility and validity. J Eval Clin Pract 2009;15:1230—5. 12. Riley RD, Hayden JA, Steyerberg EW, Moons KG, Abrams K, Kyzas PA, et al. Prognosis research strategy (PROGRESS) 2: prognosic factor research. PloS Med 2013;10(2):e1001380, http://dx.doi.org/10.1371/journal.pmed.1001380. 13. Gill CJ, Sabin L, Schmid CH. Why clinicians are natural bayesians. Brit Med J 2005;330:1080—3. 14. Rutten ALB, Stolper CF, Lugten RF, Barthels RJ. ‘Cure’ as the gold standard for likelihood ratio assessment: theoretical considerations. Homeopathy 2004;93:78—83. 15. Naranjo CA, Busto U, Sellers EM, Sandor P, Ruiz I, Roberts EA, et al. A method for estimating the probability of adverse drug reactions. Clin Pharmacol Ther 1981;80:289—95. 16. Arulmani R, Rajendran SD, Suresh B. Adverse drug reaction monitoring in a secondary care hospital in South India. Br J Clin Pharmacol 2008;65(2):210—6. 17. Rutten ALB. Comparison of effectiveness of freand infrequently used homeopathic quently medicines. Homeopathy 2011;100:175—82, http://dx.doi.org/10.1016/j.homp.2010.11.005.

The importance of case histories for accepting and improving homeopathy.

Case histories are necessary besides other types of evidence to convince doctors of a specific type action of homeopathic medicines. Prognosis of trea...
382KB Sizes 0 Downloads 0 Views