Bach Remedies research protocols paper

Papers by Walach and  Ernst. Say in both fatal flaw (or at least very serious limitation) is that they don’t do individual  diagnosis, but give identical remedy to all participants in the study.

• The Walach study particularly flawed – a mix of TEN different remedies! Aimed at ‘scattergun’ approach. Cf with Bach’s original effort, did trial of all 38 remedies combined ‘didn’t find it satisfactory’ (Most I think Dr Bach ever used was about seven, but much more usual to stick to no more than about five remedies, and ideally zeroing in to just one or two if possible). Explain key aim of diagnosis is zeroing in on the correct remedy (or combination) – not every single one potentially relevant, but the minimum essential to do this. Takes an experienced Bach practitioner to do this well.
• Ernst study somewhat more useful, as the ‘rescue remedy’ IS often recommended for ‘stress’ situations. However, also not what an experienced Bach practitioner would recommend usually. It’s designed as emergency remedy for extreme stress (analyze individual components, give example of Bach with Cromer fisherman, also your own use with post-traumatic stress – though even here you ended up focusing on just TWO of the components of the rescue remedy). A bach practitioner would also recommend a more individual consultation. Note that many of the ‘exam stress’ people would NOT probably be in the kind of extreme stress that the rescue remedy is even designed for. However, mostly the stress would be of quite different nature. And very different remedies would be indicated. E.g. hornbeam possibly relevant if procrastinating, Mimulus if fear, Elm overwhelm, quite possibly some components of the rescue remedy if shock and terror, or losing touch with present (clematis); But would need to narrow down to just a handful of these (ideally just one or two – almost always the most effective one would be a maximum of three or four). FOCUS is the key to a good remedy prescription. And protocols to test the bach remedy need to address this – otherwise it’s an “outsider” approach which isn’t following correctly the protocols of the technique itself, so not a fair test (example of testing astrology by looking at quality of predictions of newspaper ‘sun sign’ astrology columns, rather than individual astrological analyses by a highly regarded expert?). I do suggest below two possible protocols that would overcome this.
• Note the Ernst study does certainly provide useful information – that using Rescue Remedy as a ‘generic stress’ reliever is not apparently effective statistically for a large group. This is certainly how it is quite often touted – but probably not by experienced Bach practitioners! They would say – certainly better than nothing, give it a go – and WOULD quite possibly recommend if in great stress on the day itself (which would apply probably to some of the sample – but probably not most of them. Only a relatively small minority of students (in my judgement, and I have fair bit of experience of helping students through exams!) would experience the level of extreme shock and terror that the Rescue Remedy is designed for. And even in those cases, an experienced Bach practitioner would want – when time was available – to try and zero in on a more precise formulation (probably using some of the RR components, but not necessarily all five, and considering if other remedies are also relevant, e.g. Elm if overwhelmed, Gorse if sense of hopelessness etc.).
• Essential a VERY EXPERIENCED practitioner is used in any protocols to maximize effectiveness of the test. Although superficially quite simple – certainly possible for anyone to have a quick look through and make a sensible guess at a diagnosis – doing so consistently and effectively takes a fair bit of experience. Absolute minimum would be a BFRP qualification; though to my mind this wouldn’t be sufficient in itself, I would also look for a practitioner of at least 10 years experience beyond that point, and who is in addition highly regarded as a practitioner, with a successful record (e.g. as judged by the Bach Centre?). As an analogy, if looking to optimize diagnostic validity in a clinical assessment with a medical doctor, you would choose a consultant with at least ten years or so post-qualification experience, rather than a newly-qualified doctor.
• Note – in relation to the above point. The Bach system is a relatively simple one, and is designed to be used by individuals diagnosing themselves (or their friends) very often, and certainly doesn’t in normal use require a qualified Bach practitioner. The vast majority of diagnoses will be performed by people working it out for themselves (just as I have done for myself, and occasionally friends, over the last 20 years, without any formal training). However, it is relatively easy to not get the optimum remedy possible, and for conducting an experimental test, it is essential (in my view) that the most experienced practitioners possible should be used, so that “poor diagnosis” is unlikely to be the reason why the treatment is ineffective (or is much less than optimal, even if it has some limited effectiveness). Being able to ‘focus in’ on a limited group of applicable remedies, in relation to a particular person at a particular time, is a skill which certainly develops with training and experience. Distinguishing between some of the remedies can be quite tricky on occasion. There is also a tendency with inexperienced users to prescribe large numbers of remedies – any which seem potentially to have any relevance at all to the situation – which is likely to be of limited use. In my experience (echoed by Bach practitioners I have consulted with in my training) a much more focused remedy set is far more likely to yield better results. And certainly to conduct a fair and accurate test, eliminating the significant possibility of poor diagnosis is a key element.

Two possible protocols which would overcome the issues above.

• One is very simple, cheap and easy to do, and easy to replicate! Based on the idea that the Bach remedies don’t just work on humans, but on animals and indeed plants. For this I use the latter, Cress plants (bought from supermarket or grown from seed). [You should actually DO at least one run of this – ideally two or three]. This would involve putting them under severe stress by use of salt water (a major stressor for all plants except possibly some few which have adapted to cope with it, which certainly doesn’t include stress). A randomly allocated group of plants would be given Bach Rescue Remedy (incorporated into the daily salt water addition), compared with a control group given a dummy remedy [purpose of ‘dummy’ remedy is obviously not the usual one of a psychological placebo! But to ensure that the ‘carrier’ aspects of the remedy – mixture of brandy and water – are matched in terms of the drops added). This study would effectively be ‘double blind’ – obviously the plants won’t be conscious of what they are getting, but neither should the administrators of the remedy (daily) know which are which.
• [Clearly this isn’t doing the ‘individual assessment’ element which is at the heart of a normal Bach consultation, but by choosing a simple organism – cress plant – and a severe stressor, then using rescue remedy would be expected to have some positive impact on the plant adjusting to the stressor more successfully, at least that would be my hypothesis to be tested here]
• The aim is to add regular doses of salt water to produce something close to “LD50”, i.e. sufficient to effectively ‘kill off’ approximately half the plants. Preliminary testing would be needed to establish what kind of dose was needed BEFORE any experiment was run. The reason for this is to avoid ‘floor’ and ‘ceiling’ effects. If the stressor is sufficiently weak that it has very little impact on the plants, then a) it would be impossible to see if the remedy had any impact on alleviating the stress, as there wouldn’t be much to alleviate b) cf the Ernst study above, the stress needs to be SEVERE to justify use of the rescue remedy in the first place. LD50 is about right. This has been addressing ‘floor’ effects (too little salt water producing insufficient negative impact). ‘Ceiling’ effects also need to be avoided – where the salt water dose is so large it rapidly kills off 100% of the plants. The Rescue Remedy isn’t envisaged as a miracle cure in the Bach system! (E.g. to take a human equivalent, if in a severe car accident someone had had all four limbs severed and was rapidly bleeding to death, giving the rescue remedy wouldn’t in itself produce a magical cure! Though a Bach practitioner would see it having a role to play alongside the essential ‘external’ treatment). Back to the floor/ceiling effects with the Cress plants, the level of stress needs to be sufficient to kill off (or severely stunt) a good number of the plants, but not all of them. [You would expect a range of responses for any group of plants due to natural variations in precisely how they are potted and so on, even with genetically- identical plants, as is presumably the case here]. So aiming at somewhere around LD50 in preliminary trials should help produce a saltwater dose that gives scope for the Bach remedy to potentially play a role.  I would suggest a minimum of 20 plants (cress divided up into small pots with potting compost), divided into two groups of 10. The rescue remedy should be made up in advance (together with the dummy remedy – as identical as possible in water/alcohol content, simply not containing the flower remedy). The two remedies (dummy and actual) should then be added to treatment bottles labelled A and B, with a third party (not involved with the administration of the experiment) brought in to decide which of the two will become the experimental (Bach remedy) bottle, and which the control (dummy remedy). Each plant should be clearly labelled A or B. The two sets of plants should be put in a place with approximately equal lighting and heating across the whole set of plants. In addition, the two sets should be intermixed, so as to aim to cancel out any small variations which may remain (this makes it a little more difficult to administer the remedies, as they are not simply in two separate groupings, but as long as each plant is clearly labelled, and care is taken not to mix them up in administering the Remedy A/B, then this shouldn’t be a problem).
• After a sufficient period (7 days? This can also be established with preliminary testing) – the plants should then be independently rated. A scale of 1-10 perhaps (with 1 equalling dead, 10 looking completely healthy). To aid the raters, there could be some kind of photographic protocol of plants from previous trials to give examples of plants on different points on the scale, to help improve reliability. (At least) two independent raters should be used, and a Spearman rank correlation coefficient of their ratings calculated (which should be above 0.7 for reliability). It may be useful for the purposes of this rating that the raters are not involved in the design and running of the experiment, and possibly also for the A/B designation on each plant to be temporarily covered up or obscured while the rating is being done (to try and avoid even a subconscious sense of which group is doing ‘better’ or ‘worse’ – though of course the raters, along with the experimenters, won’t at this stage know which of the A or B groupings is the one receiving the flower remedy). Then the mean scores for each group should be calculated – the first thing obviously to check is whether the Bach group does indeed have a higher mean rating (i.e. appearing healthier) than the control group. If the control group is higher, then the hypothesis fails at this point. If the Bach grouping is indeed higher, then it is necessary to calculate whether the difference is statistically significant (at least at the p-value less than 0.05 level). [Appropriate tests – t-test? Non-parametric? Criteria for choosing?]
• Note with both of these protocols, although I have brought up every aspect I could think of to produce a better designed and controlled experiment, which will hopefully be useful, others may well be able to refine these designs further. Although I have a basic understanding of experimental design (from my time doing research in experimental physics, and subsequently through my training as a psychologist), I am these days more familiar with qualitative research than experimental. So no doubt someone more familiar with the statistical and experimental issues would be able to develop things further.
• I do think though that a lack of understanding of how Bach remedies are prescribed has been a flaw behind most empirical designs so far – the two main studies reviewed both give the identical remedy to all participants. While there is on occasion some scope to do this with the rescue remedy (as in my ‘cress’ experimental protocol), particularly in with human participants a much more individually-targeted approach would be needed. I suggest a possible protocol below.
• This could be based on ‘exam stress’ (similar to the Harald Walach study above), but would avoid the flaw in this study that involved administering the identical ten remedies to all participants. I should say the ten remedies in the Walach study were sensibly chosen, in that these were remedies which would very often be potentially relevant for a  Bach practitioner (or individual, if self-diagnosing) to consider if looking at stress. However, firstly, a Bach practitioner would almost never administer this many remedies – the key principle in diagnosis is to focus it down on as few as possible, usually two or three, or perhaps even a single remedy. Sometimes four or five, very rarely more than this, and I’ve never heard of ten being given. And even in the case where a relatively large number of remedies are given (say five), this would ALWAYS involve an individual assessment (either by a practitioner or the individual themselves, if sufficiently knowledgeable of the Bach system). The one exception is of course the ‘Rescue Remedy’ combination of five individual remedies, which is generic. But although probably the best known of the Bach remedies, and a useful ‘first step’ in a crisis situation, it is definitely not a substitute for a more considered diagnostic assessment once the time and situation allows for this.
• So in my revised protocol, much would remain the same as with the Walach study. However, instead of a ‘ten remedy’ selection given identically to all participants, they would be assessed by an experienced Bach Remedies practitioner [see above]. This practitioner would assess firstly if they were actually experiencing a significant amount of nerves by discussing it with the individual – any who didn’t appear to be doing so could be excluded from further participation in the study (as there is no scope to improve the nerves if there is none, or very little, there to start with). Those remaining would then be prescribed the appropriate remedy. I do think it important that they tried to take it on a regular basis – four times suggested daily (the Walach study, although aiming at this, apparently was ok if it was only taken once per day. Of course this is understandable, given that it was important to maximize the number of participants, and that they may not all be organized enough to take it four times. However, this does weaken the study quite a lot if it is only once per day, so you would want to try hard to standardize this, and check compliance perhaps with a checkbox participants filled in or something similar).
• This is the key part of the protocol. ALL participants – both the experimental group (receiving the Bach remedy) and the control group (receiving a placebo) would need to have the assessment. Otherwise there is of course a major confounding variable (in that the assessment itself could certainly be seen to contain a ‘counselling’ element which would in itself help reduce stress and be of therapeutic value, quite apart from any subsequent administration of Bach remedy). But only half of the group – randomly allocated – would be given the actual remedy, the rest would be given a placebo. (The decision and administration of this should not be done by the Bach practitioner, but by some individual who is not otherwise involved in the experimental team, to ensure lack of knowledge of which one is which in subsequent administration of the experiment, until the final results are being collated).
• This should be done at least 10 days before the exam, ideally earlier, to allow the remedies to get to work.
• There would then need to be some assessment of ‘effectiveness’. Ideally multiple methods would be used. Physiological ones may be possible? (Measures of skin resistance etc. as markers of stress, on the day of the exam, or possibly the day before). Some kind of standard stress questionnaire? A self-report rating of how they felt they had benefited from the remedies (it is quite likely that a majority of the participants, both control and experimental groups, will report some benefit – both because of placebo effect, and the potential ‘counselling’ element in the Bach consultation itself, which all would have received. The hypothesis under examination of course is whether those receiving the actual Bach treatment show some additional statistically-significant benefit). Another measure could be of some measure of exam effectiveness. It would be insufficient to just take the raw exam scores as there would be great variability in terms of individual ability. However perhaps there could be an estimated exam score by the teacher for each student, based on their performance on continuous assessment during the module. And the factor being examined would therefore be predicted score minus actual score. Inference being here that reduced exam nerves should result in a bigger figure for this factor (as students with less anxiety can reasonably be expected to reach or exceed their potential more than those with greater anxiety). These two measures – self-report scores, and “predicted minus actual” exam scores –are relatively easy to administer and calculate, so are worth doing.
• In addition to these two measures, some measure of emotional response should also be taken. From the point of view of the Bach system itself, a more harmonious emotional state would actually be seen as the primary aim. Of course one would expect that this would also tend to lead to better performance (in all sorts of ways), but the harmonious emotional state would be seen as an extremely valuable end in itself, given the spiritual focus of the Bach system on enhancing the life quality of each person. Although not easy to measure, some kind of questionnaire relating to designed to assess quality of life could be administered both before and after the Bach treatments, to see if there was a significant difference in any changes produced between the control and Bach group.
• The writings of Edward Bach himself tended to decry research, saying essentially that the system ‘proved itself’ in use. Certainly the millions of people (?) certainly hundreds of thousands, who use this on a regular basis feel it is beneficial. Though to the conventionally scientific mind, there is sufficient capacity for human self-deception to find this unsatisfactory. I do think there is however scope for ‘case studies’ to still make a useful contribution to the research evidence base for Bach studies. Placebo effect is always going to be a very significant potential issue – although the potential to adminster it to very young children, animals, and plants would potentially obviate this aspect. (The absence of a control in such a case study would of course still be a significant issue, but a build up of significant case study evidence is nonetheless of value in assessing whether a more carefully-designed research study is worth undertaking).
• Hopefully the critiques and constructive suggestions I’ve provided in this paper will help any future researchers who wish to examine the efficacy of Bach flower remedies.