“Working well for him…” That was the response to my tweet yesterday about cupping for Olympic swimmers. I had tweeted this picture showing one swimmer’s cupping marks (similar signs currently are currently being displayed by several competitors in Tokyo).
I had added to the tweet my post from 2018 which failed to show that cupping is an effective means of improving athletic performance.
The response ‘WORKING WELL FOR HIM..’ irritated me (not that it has the slightest importance) and made me think how prone we all are to find causal relationships where there are, in fact, none (which might have more importance). I feel that we must, as intelligent humans, do more to fight this reflex.
In 2008, just before Simon Singh and I published ‘TRICK OT TREATMENT?‘, I broke my left shoulder. It was stupid, painful, unpleasant, and most annoying. Yet, it coincided with a very nice publishing success: our book received plenty of praise and was translated into about 20 languages.
So, should we recommend to all authors who are about to publish a book that they break their left shoulder? I think we can probably agree that this would be absurd.
But why do many people who see the cupping-marked Olympic athletes think that cupping is WORKING WELL FOR THEM? I know, it is tempting to think that they know best, and they must have tested it, etc. But why not rather consult the evidence? Why not rather question the plausibility of cupping as a means to improve performance? Why not rather consider that athletes do all sorts of weird, irrational things that make them feel a little more secure?
Frankly, the evidence that breaking your arm makes you publish a decent book is just as sound as the evidence that cupping improves the speed of swimmers. My advice, therefore, is to resist quick thinking where slow thinking including asking probing questions and consulting the evidence is indicated.
Chinese researchers evaluated the effect of Chinese medicine (CM) on survival time and quality of life (QoL) in patients with small-cell lung cancer (SCLC). They conducted an exploratory and prospective clinical observation. Patients diagnosed with SCLC receiving CM treatment as an add-on to conventional cancer therapies were included and followed up every 3 months. The primary outcome was overall survival (OS), and the secondary outcomes were progression-free survival (PFS) and QoL.
A total of 136 patients including 65 limited-stage SCLC (LS-SCLC) patients and 71 extensive-stage SCLC (ES-SCLC) patients were analyzed. The median OS of ES-SCLC patients was 17.27 months, and the median OS of LS-SCLC was 40.07 months. The survival time was 16.27 months for SCLC patients with brain metastasis, 9.83 months for liver metastasis, 13.43 months for bone metastasis, and 18.13 months for lung metastasis. Advanced age, pleural fluid, liver, and brain metastasis were risk factors, while longer CM treatment duration was a protective factor. QoL assessment indicated that after 6 months of CM treatment, scores increased in function domains and decreased in symptom domains.
The authors concluded that CM treatment might help prolong OS of SCLC patients. Moreover, CM treatment brought the trend of symptom amelioration and QoL improvement. These results provide preliminary evidence for applying CM in SCLC multi-disciplinary treatment.
Sorry, but these results provide NO evidence for applying CM in SCLC multi-disciplinary treatment! Even if the findings were a bit better than those reported for SCLC in the literature – and I am not sure they are – it is simply not possible to say with any degree of certainty what effect the CM had. For that, we would obviously need a proper control group.
The study was supported by the National Natural Science Foundation of China (No. 81673797), and Beijing Municipal Natural Science Foundation (No. 7182142). In my view, this paper is an example for showing how the relentless promotion of dubious Traditional Chinese Medicine by Chinese officials might cost lives.
I feel that it is time to do something about it.
But what precisely?
Any ideas anyone?
Recently, I received this comment from a reader:
Edzard-‘I see you do not understand much of trial design’ is true BUT I wager that you are in the same boat when it comes to a design of a trial for LBP treatment: not only you but many other therapists. There are too many variables in the treatment relationship that would allow genuine , valid criticism of any design. If I have to pick one book of the several listed elsewhere I choose Gregory Grieve’s ‘Common Vertebral Joint Problems’. Get it, read it, think about it and with sufficient luck you may come to realize that your warranted prejudices against many unconventional ‘medical’ treatments should not be of the same strength when it comes to judging the physical therapy of some spinal problems as described in the book.
And a chiro added:
EE: I see that you do not understand much of trial design
Perhaps it’s Ernst who doesnt understand how to research back pain.
“The identification of patient subgroups that respond best to specific interventions has been set as a key priority in LBP research for the past 2 decades.2,7 In parallel, surveys of clinicians managing LBP show that there are strong views against generic treatment and an expectation that treatment should be individualized to the patient.6,22.”
Journal of Orthopaedic & Sports Physical Therapy
Published Online:January 31, 2017Volume47Issue2Pages44-48
Do I need to explain why the Grieve book (yes, I have it and yes, I read it) is not a substitute for evidence that an intervention or technique is effective? No, I didn’t think so. This needs to come from a decent clinical trial.
And how would one design a trial of LBP (low back pain) that would be a meaningful first step and account for the “many variables in the treatment relationship”?
How about proceeding as follows (the steps are not necessarily in that order):
- Study the previously published literature.
- Talk to other experts.
- Recruit a research team that covers all the expertise you need (and don’t have yourself).
- Formulate your research question. Mine would be IS THERAPY XY MORE EFFECTIVE THAN USUAL CARE FOR CHRONIC LBP? I know LBP is but a vague symptom. This does, however, not necessarily matter (see below).
- Define primary and secondary outcome measures, e.g. pain, QoL, function, as well as the validated methods with which they will be quantified.
- Clarify the method you employ for monitoring adverse effects.
- Do a small pilot study.
- Involve a statistician.
- Calculate the required sample size of your study.
- Consider going multi-center with your trial if you are short of patients.
- Define chronic LBP as closely as you can. If there is evidence that a certain type of patient responds better to the therapy xy than others, that might be considered in the definition of the type of LBP.
- List all inclusion and exclusion criteria.
- Make sure you include randomization in the design.
- Randomization should be to groups A and B. Group A receives treatment xy, while group B receives usual care.
- Write down what A and B should and should not entail.
- Make sure you include blinding of the outcome assessors and data evaluators.
- Define how frequently the treatments should be administered and for how long.
- Make sure all therapists employed in the study are of a high standard and define the criteria of this standard.
- Train all therapists of both groups such that they provide treatments that are as uniform as possible.
- Work out a reasonable statistical plan for evaluating the results.
- Write all this down in a protocol.
Such a trial design does not need patient or therapist blinding nor does it require a placebo. The information it would provide is, of course, limited in several ways. Yet it would be a rigorous test of the research question.
If the results of the study are positive, one might consider thinking of an adequate sham treatment to match therapy xy and of other ways of firming up the evidence.
As LBP is not a disease but a symptom, the study does not aim to include patients that all are equal in all aspects of their condition. If some patients turn out to respond better than others, one can later check whether they have identifiable characteristics. Subsequently, one would need to do a trial to test whether the assumption is true.
Therapy xy is complex and needs to be tailored to the characteristics of each patient? That is not necessarily an unsolvable problem. Within limits, it is possible to allow each therapist the freedom to chose the approach he/she thinks is optimal. If the freedom needed is considerable, this might change the research question to something like ‘IS THAT TYPE OF THERAPIST MORE EFFECTIVE THAN THOSE EMPLOYING USUAL CARE FOR CHRONIC LBP?’
My trial would obviously not answer all the open questions. Yet it would be a reasonable start for evaluating a therapy that has not yet been submitted to clinical trials. Subsequent trials could build on its results.
I am sure that I have forgotten lots of details. If they come up in discussion, I can try to incorporate them into the study design.
This study assessed the effects of being born under the zodiac sign Pisces on mortality. For that purpose, a retrospective observational study was conducted of the data from 26 Scandinavian intensive care units between 2009 and 2011. Patients aged 18 years or older with severe sepsis and in need of fluid resuscitation were included from the Scandinavian Starch for Severe Sepsis/ Septic Shock (6S) trial. The main outcome measure was the 90-day mortality.
The researchers included all 798 patients in the study; 70 (9%) of them were born under the sign of Pisces. The primary outcome (death within 90 days) occurred in 25 patients (35.7%) in the Pisces group, compared with 348 patients (48%) in the non-Pisces group (relative risk, 0.75; 95% CI, 0.54-1.03; one-sided P = 0.03).
The authors concluded that in a multicentre randomised clinical trial of IV fluids, being born under the sign of Pisces was associated with a decreased risk of death. Our study shows that with convenient use of statistics and an enticing explanatory hypothesis, it is possible to achieve significant findings in post-hoc analyses of data from large trials.
This is an excellent paper! It showcases the sort of nonsense one can do with datasets, statistics, and post hoc hypotheses. The authors entitled their article “Gone fishing in a fluid trial”, and this title should ensure that there are not some astrology nutters who mistake correlation for causation
… I hope
… but, of course, I am an optimist.
Homeopathy is sometimes claimed to be effective for primary dysmenorrhoea (PD), but the claim is not supported by sound evidence. This study was undertaken to examine the efficacy of individualized homeopathic medicines (IH) against placebo in the treatment of PD.
A double-blind, randomized, placebo-controlled trial was conducted at the gynecology outpatient department of Mahesh Bhattacharyya Homoeopathic Medical College and Hospital, West Bengal, India. Patients were randomized to receive either IH (n=64) or identical-looking placebo (n=64). Primary and secondary outcome measures were 0-10 numeric rating scales (NRS) measuring the intensity of pain of dysmenorrhea and verbal multidimensional scoring system (VMSS) respectively, all measured at baseline, and every month, up to 3 months.
The two groups were comparable at baseline. The attrition rate was 10.9% (IH: 7, placebo: 7). Differences between groups in both pain NRS and VMSS favored IH over placebo at all time points with medium to large effect sizes. Natrum muriaticum and Pulsatilla nigricans were the most frequently prescribed medicines. No harms, serious adverse events, or intercurrent illnesses were recorded in either group.
The authors concluded that homeopathic medicines acted significantly better than placebo in the treatment of PD. Independent replication is warranted.
A previously published RCT could not show any significant effect of homeopathy on primary dysmenorrhea in comparison with placebo. The authors of the new study claim that the discrepant findings might be due to the fact that IH requires great skill. In other words, negative studies are according to this explanation negative not because homeopathy does not work but because the prescribers are not up to it. Such notions have often been voiced on this blog and elsewhere and are used as a veritable ‘get-out clause’ for homeopathy: ONLY THE POSITIVE RESULTS ARE VALID! Consequently, systematic reviews of the evidence must only consider positive trials. And this, of course, means that the findings are invariable positive.
I find this more than a little naive and would much prefer to wait for an independent replication where ‘independent’ means that the trial is run by experts who are not advocates of homeopathy (as in the present trial).
Bach flower remedies were invented in the 1920s by Dr. Edward Bach (1886-1936), a doctor homeopath who had previously worked in the London Homeopathic Hospital. They have since become very popular in Europe and beyond. Bach flower remedies are clearly inspired by homeopathy; however, they are not the same because they do not follow the ‘like cures like’ principle and are they potentized. They are manufactured by placing freshly picked specific flowers or parts of plants in water which is subsequently mixed with alcohol, bottled, and sold. Like most homeopathic remedies, they are highly dilute and thus do not contain therapeutic amounts of the plant printed on the bottle.
The aim of this new randomized, double-blind, placebo-controlled trial was to compare the efficacy of flower therapy for the treatment of anxiety in overweight or obese adults with that of a placebo. The authors examined improvement in sleep patterns, reduction in binge eating, and change in resting heart rate (RHR).
The study included 40 participants in the placebo group and 41 in the intervention group. Participants were of both genders, from 20 to 59 years of age, overweight or obese, with moderate to high anxiety. They were randomized into two groups:
- one group was treated with Bach flower remedies (BFR) (bottles containing 30 mL of 30% hydro-brandy solution with two drops each of Impatiens, White Chestnut, Cherry Plum, Chicory, Crab Apple, and Pine), purchased from Healing® Flower Essences (São Paulo, Brazil)
- the other group was given a placebo (same solution without BFR).
All patients were instructed to orally ingest the solutions by placing four drops directly in the mouth four times a day for 4 weeks.
The primary outcome was anxiety (State-Trait Anxiety Inventory [STAI]). Secondary outcomes were sleep (Pittsburgh Sleep Quality Index [PSQI]), binge eating (Binge Eating Scale [BES]), and RHR (electrocardiogram).
Multivariate analysis showed significant reductions in scores for the following variables in the intervention group when compared with the placebo group: STAI (β = −0.190; p < 0.001), PSQI (β = −0.160; p = 0.027), BES (β = −0.226; p = 0.001), and RHR (β = −0.07; p = 0.003).
The authors concluded that anxiety symptoms, binge eating, and RHRs of the individuals treated with flower therapy decreased, and their sleep patterns improved when compared with those treated with the placebo.
Did the alcohol in the verum preparation had a relaxing effect? No, I was teasing. The amount would have been too small and the effect would have been the same in both groups. But what could have caused the observed outcome? I have to admit that I have no idea.
I read the study several times and could not find a major flaw. Hence it must have been the flower remedy that caused the positive outcome? No, I am teasing again. I find this impossible to imagine. These remedies contain nothing that might explain the results and all previous systematic reviews of all the available trials have all reached a negative conclusion. Before I seriously consider the option that flower remedies are more than placebos, I would like to see an independent replication.
The integration of so-called alternative medicine (SCAM) into cancer care may reduce the adverse effects of anti-neoplastic treatment but also cause new problems and non-adherence to conventional treatment. Therefore, its net benefit is questionable.
The aim of this randomized controlled study was to investigate the impact of integrative open dialogue about SCAM on cancer patients’ health and quality of life (QoL).
Patients undergoing curative or palliative anti-neoplastic treatment were randomly assigned to standard care (SC) plus SCAM or SC alone. A nurse specialist facilitated SCAM in one or two sessions. The primary endpoint was the
frequency of grade 3–4 adverse events (AE) eight weeks after enrollment. Secondary endpoints were the frequency of grade 1–4 AE and patient-reported QoL, psychological distress, perceived information, attitude towards and use of SCAM 12 and 24 weeks after enrollment. Survival was analyzed post-hoc.
Fifty-seven patients were randomized to SCAM and 55 to SC. No significant differences were found in terms of AEs of cancer patients. A trend towards better QoL, improved survival, and a lower level of anxiety was found in the SCAM group.
The authors concluded that integration of SCAM into daily oncology care is feasible. IOD-CAM was not superior to SC in reducing the frequency of grade 3-4 AEs, but it did not compromise patient safety. Implementation of SCAM
may improve the QoL, anxiety, and emotional well-being of the patients by reducing the level of nausea, vomiting and diarrhea. Finally, SCAM potentially improves the patients’ self-care, which contributes to
increased treatment adherence and improved survival.
This is an interesting paper with a very odd conclusion. The positive trends found failed to be statistically significant. Why employ statistics only to ignore them in our interpretation of the findings?
I can well imagine that the integration of effective treatments into cancer care improves the outcome. I have no problem with this at all – except it is not called INTEGRATIVE MEDICINE but EVIDENCE-BASED MEDICINE!!! If we integrate dubious treatments into cancer care, it’s called INTEGRATIVE MEDICINE, and it’s unlikely to do any good.
In my view, this small study showed just one thing:
Integrative medicine does not reduce adverse effects in cancer patients.
Post-traumatic stress disorder (PTSD), previously known as battle fatigue syndrome or shell shock, is a condition that can be triggered by the experience of some frightening event. PTSD can be debilitating leading to the production of feelings of helplessness, intense fear, and horror. Numerous treatments of PTSD exist but few have been shown to be truly effective. A team of Canadian researchers explored the effects of cannabis on PTSD symptoms, quality of life (QOL), and return to work (RTW). Their systematic review also investigated harms such as adverse effects and dropouts due to adverse effects, inefficacy, and all-cause dropout rates.
Their electronic searches located one RCT and 10 observational studies (n = 4672). Risk of bias (RoB) was assessed with the Cochrane risk of bias tool and ROBINS-I. Evidence from the included studies was mainly based on studies with no comparators. Results from unpooled, high RoB studies suggested that cannabis was associated with a reduction in overall PTSD symptoms and improved QOL. Dry mouth, headaches, and psychoactive effects such as agitation and euphoria were the most commonly reported adverse effects. In most studies, cannabis was well tolerated. A small proportion of patients experienced a worsening of PTSD symptoms.
The authors concluded that the evidence in the current study primarily stems from low quality and high RoB observational studies. Further RCTs investigating cannabis effects on PTSD treatment should be conducted with larger sample sizes and explore a broader range of patient-important outcomes.
Various drugs are currently used for the treatment of PTSD including selective serotonin reuptake inhibitors; tricyclic antidepressants (amitriptyline and isocarboxazid); mood stabilizers (Divalproex and lamotrigine); atypical antipsychotics (aripiprazole and quetiapine) but their effectiveness has not been proven. A recent systematic review included 30 RCTs of a range of heterogeneous non-psychological and non-pharmacological interventions. There was emerging evidence for 6 different approaches:
- saikokeishikankyoto (a herbal preparation),
- somatic experiencing,
- transcranial magnetic stimulation,
This list makes me wonder: are these treatments, including cannabis, truly promising, or is PTSD one of those conditions for which nearly every treatment works a little because of its placebo effect?
Due to polypharmacy and the rising popularity of so-called alternative medicines (SCAM), oncology patients are particularly at risk of drug-drug interactions (DDI) or herb-drug interactions (HDI). The aims of this study were to assess DDI and HDI in outpatients taking oral anticancer drugs.
All prescribed and non-prescribed medications, including SCAMs, were prospectively collected by hospital pharmacists during a structured interview with the patient. DDI and HDI were analyzed using four interaction software programs: Thériaque®, Drugs.com®, Hédrine, and Memorial Sloan Kettering Cancer Center (MSKCC) database. All detected interactions were characterized by severity, risk, and action mechanism. The need for pharmaceutical intervention to modify drug use was determined on a case-by-case basis.
A total of 294 patients were included, with a mean age of 67 years [55-79]. The median number of chronic drugs per patient was 8 [1-29] and 55% of patients used at least one SCAM. At least 1 interaction was found for 267 patients (90.8%): 263 (89.4%) with DDI, 68 (23.1%) with HDI, and 64 (21.7%) with both DDI and HDI. Only 13% of the DDI were found in Thériaque® and Drugs.com® databases, and 125 (2.5%) were reported with a similar level of risk on both databases. 104 HDI were identified with only 9.5% of the interactions found in both databases. 103 pharmaceutical interventions were performed, involving 61 patients (20.7%).
The authors concluded that potentially clinically relevant drug interactions were frequently identified in this study, showing that several databases and structured screening are required to detect more interactions and optimize medication safety.
These data imply that DDIs are more frequent than HDIs. This does, however, not tell us which are more important. One crucial difference between DDIs and HDIs is that the former are usually known to the oncology team who should thus be able to prevent them or deal with them appropriately; in contrast, HDIs are often not known to the oncology team because many patients fail to disclose the fact that they take herbal remedies. Some forget, some do not think of herbals as medicine, others may be worried about their physician’s reaction.
It follows that firstly, conventional healthcare practitioners should always ask about the usage of herbal remedies, and secondly, they need to be informed about which herbal remedy might interact with which drug. The first can easily be implemented into routine history-taking; the second is more problematic, not least because our knowledge about HDIs is still woefully incomplete. In view of this, it might often be wise to tell patients to stop taking herbal remedies while they are on prescription drugs.
Vertebral artery dissections (VAD) are a rare but important cause of ischemic stroke, especially in younger patients. Many etiologies have been identified, including motor vehicle accidents, cervical fractures, falls, physical exercise, and, as I have often discussed on this blog, cervical chiropractic manipulation. The goal of this study was to investigate the subgroup of patients who suffered a chiropractor-associated injury and determine how their prognosis compared to other-cause VAD.
The researchers, neurosurgeons from Chicago, conducted a retrospective chart review of 310 patients with vertebral artery dissections who presented at their institution between January 2004 and December 2018. Variables included demographic data, event characteristics, treatment, radiographic outcomes, and clinical outcomes measured using the modified Rankin Scale.
Overall, 34 out of our 310 patients suffered a chiropractor-associated injury. These patients tended to be younger (p = 0.01), female (p = 0.003), and have fewer comorbidities (p = 0.005) compared to patients with other-cause VADs. The characteristics of the injuries were similar, but chiropractor-associated injuries appeared to be milder at discharge and at follow-up. A higher proportion of the chiropractor-associated group had injuries in the 0-2 mRS range at discharge and at 3 months (p = 0.05, p = 0.04) and no patients suffered severe long-term neurologic consequences or death (0% vs. 9.8%, p = 0.05). However, when a multivariate binomial regression was performed, these effects dissipated and the only independent predictor of a worse injury at discharge was the presence of a cervical spine fracture (p < 0.001).
The authors concluded that chiropractor-associated injuries are similar to VADs of other causes, and apparent differences in the severity of the injury are likely due to demographic differences between the two populations.
The authors of the present paper are clear: “chiropractic manipulations are a risk factor for vertebral artery dissections.” This fact is further supported by a host of other investigations. For instance, the Canadian Stroke Consortium found that 28% of strokes following cervical artery dissection were preceded by chiropractic neck manipulation. Dziewas et al. obtained a similar rate in patients with vertebral artery dissections. Many chiropractors are in denial; however, this is merely due to their overt conflicts of interest.
My conclusions from the accumulated evidence are this:
Spinal manipulations of the upper spine should not be routinely used for any condition. Patients who nevertheless insist on having them must be made aware of the risks and give informed consent.